12016June

7

Eye-tracking analysisof face observing and face recognition

Andrej Iskra, Helena Gabrijelčič Tomc

The analysis of ink jetprinted eco-font efficiency

Rastko Milošević, Uroš Nedeljković,Bojan Banjanin, Dragoljub Novaković,

Nemanja Kašiković

Development of tactile floor planfor the blind and the visually impaired

by 3D printing techniqueRaša Urbas, Matej Pivar, Urška Stankovič Elesini

Simulation of durability the evaluation ofrelief-dot Braille images

Svitlana Havenko, Mykola Havenko,Marta Labetska

Published by

Un�vers�ty of nov� sad, serb�afaculty of technical sciencesdepartment of Graphic engineering and design

1/2016volume 7, number 1, June 2016.

JoUrnaL of GraPH�CenG�neer�nG and des�Gn

Published by

Address:faculty of technical sciences, department of Graphic engineering and design, trg dositeja obradovića 621000 novi sad, serbia

Telephone numbers:+381 21 485 26 20+381 21 485 26 26+381 21 485 26 21

Fax number:+381 21 485 25 45

Email: jged@uns.ac.rs

Web address: www.grid.uns.ac.rs/jged

Frequency: 2 issues per yearPrinting: Faculty of Technical Sciences,Department of Graphic Engineering and Design

Circulation: 200

Electronic version of journal available onwww.grid.uns.ac.rs/jged

E-ISSN 2217-9860

the journal is abstracted/indexed in the directory of open access Journals

editors

Dragoljub Novaković, University of Novi Sad, Novi Sad, Serbia

Nemanja Kašiković, University of Novi Sad, Novi Sad, Serbia

editorial boardThomas Hoffmann-WalbeckHDM Stuttgart, Stuttgart, Germany rafael huertasUniversity of Granada, Granada, Spain Joanna ewa izdebskaWarsaw University of Technology, Warsaw, Poland Igor MajnarićUniversity of Zagreb, Zagreb, CroatiaBranko MilosavljevićUniversity of Novi Sad, Novi Sad, SerbiaTadeja MuckUniversity of Ljubljana, Ljubljana, Slovenia erzsébet NovotnyPNYME, Budapest, Hungary Anastasios PolitisTechnological Educational Institute of Athens Graphic Arts Technology, Athens, Greece Miljana PricaUniversity of Novi Sad, Novi Sad, Serbia iskren spiridonovUniversity of Chemical Technology and Metallurgy, Sofia, Bulgaria Mladen StančićUniversity of Banja Luka, Banja Luka, Bosnia and Herzegovina tomáš syrovýUniversity of Pardubice, Pardubice, Czech Republic Gojko VladićUniversity of Novi Sad, Novi Sad, Serbiathomas sabuMahatma Gandhi University, Kottayam, IndiaJonas MalinauskasVilnius College of Technologies and Design, Vilnius, LithuaniaRoberto PašićUKLO University St. Climent Ohridski, Bitola, Macedonia

C�P - Katalogizacija u publikacijibiblioteka Matice srpske, novi sad655JGed : Journal of Graphic engineering and design /editor dragoljub novaković. - vol. 1, no. 1 (nov. 2010) -sciences, department of Graphic engineering and design,2010-. 30 cmdva puta godišnje�ssn 2217-379XCob�ss.sr-�d 257662727

Art DirectorUroš Nedeljkovićlayout designBojan BanjaninJournal cover designDunja Branovački, Jelena Vladušić, Irma Puškarević

University of novi sadfaculty of technical sciences

dePartMent of GraPH�C enG�neer�nG and des�Gn

Eye-tracking analysis of face observing and face recognitionAndrej Iskra, Helena Gabrijelčič Tomc

the analysis of ink jet printed eco-font efficiencyRastko Milošević, Uroš Nedeljković, Bojan Banjanin, Dragoljub Novaković, Nemanja Kašiković

Development of tactile floor plan for the blind and the visually impaired by 3d printing techniqueRaša Urbas, Matej Pivar, Urška Stankovič Elesini

Simulation of durability the evaluation of relief-dot Braille imagesSvitlana Havenko, Mykola Havenko, Marta Labetska

We are presenting the new 12th issue of the Journal of Graphic Engineering and Design. This issue announces new editorial and reviewer board which further improves the journal with the goal of reaching the standards of leading scientific journals.

Theme of the papers published in this issue are diverse and attention worthy, because this issue covers topics from different research areas, showing wide nature of this scientific field.

Eye-tracking analysis of face observing and face recognition is always an interesting research subject.

The analysis of ink jet printed eco-font efficiency showed that it is possible to save the substantial amount of toner using this font type.

Development of tactile floor plan for the blind and the visually impaired by 3D printing technique showed us that 3D printing technique is absolutely appropriate for printing tactile floor plans.

In simulation of durability the evaluation of relief-dot Braille images authors presented the model of factors that influence the Braille durability with the definition of the hierarchy between them, which makes it possible to perform forecasting and improve operational stability of relief-dot marking.

As always we are grateful to all authors for their contribution and constant improvement of the journal’s quality.

We invite all authors interested in publishing their research to submit their papers and to join forces with us in order to improve the research field.

Dragoljub NovakovićNemanja Kašiković

dear readers,

Contents

13

19

27

5

Journal of Graphic Engineering and Design, Volume 7 (1), 2016. 5

Introduction

During development of internet various forms of infor-mation came along. Now days we have all sorts of ele-ment as web content. On basic level we can divide them as text and multimedia. Further on multimedia divides in:

• Images• Animation• Sound• Video (Kyrnin, 2014)

There is a great deal of debate why is important to include images in presentation (either web presenta-tion or publication presentation) (Patel, 2014). From all of these elements images carry most information. That also applies for face images. Web presentations of companies, institutes, universities usually also present employees. From the point of value of person presen-

tation, it is very important to include person’s image beside all other personal and professional information. There are some general factors about face images to be attractive to observers (Nielsen & Pernice, 2010):

• Smiling face• Facing the camera• Authentic• Simple background

Our research was done with eye tracking system. In general, an eye tracker is a device for measuring eye positions and eye movement. From this explanation two elements of eye tracking are defined: fixations and saccades. Fixation is when the eye gaze pauses in a cer-tain position, whereas saccades are quick, simultaneous movements of both eyes in the same direction (Cassin & Solomon, 1990). Series of fixations and saccades is called scanpath. Almost all visual information during

Andrej Iskra, Helena Gabrijelčič Tomc

University of Ljubljana, Faculty of Natural Sciences and Engineering, Snežniška 5, 1000 Ljubljana, Slovenia Corresponding author: Andrej Iskrae-mail: andrej.iskra@ntf.uni-lj.si

First recieved: 16.12.2015Accepted: 20.01.2016.

abstract

Images are one of the key elements of the content of the World Wide Web. One group of web images are also photos of people. When various institutions (universities, research organizations, companies, associa-tions, etc.) present their staff, they should include photos of people for the purpose of more informative presentation. The fact is, that there are many specifies how people see face images and how do they remem-ber them. Several methods to investigate person’s behavior during use of web content can be performed and one of the most reliable method among them is eye tracking. It is very common technique, particular-ly when it comes to observing web images. Our research focused on behavior of observing face images in process of memorizing them. Test participants were presented with face images shown at different time scale. We focused on three main face elements: eyes, mouth and nose. The results of our analysis can help not only in web presentation, which are, in principle, not limited by time observation, but especial-ly in public presentations (conferences, symposia, and meetings).

KEy WoRDSEye tracking, face recognition, fixation duration, face elements, area of interest, heat maps

eye-tracking analysis of face observing and face recognition

http://doi.org/10.24867/JGED-2016-1-005Original scientific paper

6

saccades is made in the central two degrees of the visual angle (fovea). Outside of these areas (periphery vision) is less informative. Therefor the location of fixa-tion provides us with information of location which was observed and processed during observation session.

On average, fixations last for around 200 ms during the reading of linguistic text, and 350 ms during the viewing of a scene. There are many researches according to the meaning of fixation duration. For example, in terms or reading longer fixation normally means difficult words (Pollatsek, Rayner & Balota, 1986), in terms of visual aspect refers to difficulty in extracting information (Hooge & Erkelens, 1998). At viewing face image, we deal with longer fixation duration than other image content (e. g. natural scene) (Guo et al., 2006). And which face elements attract most attention? Several authors have been exploring these most attracted face elements, which are also most important for face remembering. All researches have shown three major face elements that significantly stand out of others: eyes, mouth and nose (Henderson et al., 2001; Buchan, Pare & Munhall, 2007). Based on those three main face elements a comparison between free and restricted viewing learning condition (Henderson, Williams & Falk, 2005) was conducted.

There are also slight differences at observed time distribution between main three elements in depen-dence of face expression, gender and age (Cangöz et al., 2013). Heisz & Shore (2008) investigated how face elements scanning changes for familiar faces. Face elements are also very important in cross-face recogni-tion (stimuli and test participants from different races) (Goldinger, He & Papesh, 2009; Hills & Pake, 2013).

We conducted two tests. In the first test the aim of research was to investigate how participants viewed main three face elements according to the available time and what is the distribution of observation time for eyes, mouth and nose. Are eyes really mostly the first visited face element? How long the face presentation time must be that observer return to the eyes again?

The purpose of second test was to get basic information how well observers remembered faces according to the time of face shown. We simply measure the percent-age of correct recognition of previously shown faces.

Material and methods

Participants

Participants were recruited among students at Fac-ulty of Natural Sciences and Engineering (Ljubljana, Slovenia). They were divided into three test groups (A, B and C), so each has 14 participants. They all have normal or corrected-to-normal vision. Age range was

from 19 to 22 years old. They all volunteered in return for a small bonus at laboratory practice grade.

apparatus

As mentioned above, the key instrument for our investigation was eye tracking system. Eye tracking is mainly used for researches in the field of usability investigations in all sorts of research areas. That can be done for web sites, newspapers, magazines, games, stores (real stores in shopping centers), TV commer-cials, traffic, education, etc. (Horsley et al., 2014).

Tobii X120 eye tracking system was used in our exper-iment. It is stand-alone eye tracking unit designed for eye tracking studies of real-world flat surfaces. We used it in combination with computer’s LCD monitor.

This system has sampling rate of 120 Hz for recording eye movements. The defaults setting for definition of fixation was 100 ms for 30 pixel area. That means if gaze stayed in the area 30 pixel for at least 100 ms it was concerned as one fixation (Goldinger, He & Papesh, 2009). If gaze left the region and returned within 100 ms, it was considered to be the same gaze. Recom-mended distance form eye tracking system is 60 cm (Goldinger, He & Papesh, 2009). All recordings and analyses were done using Tobii Studio 3.1 software.

» Figure 1: Eyetracking testing setup

Stimuli

We took and downloaded 60 high resolution imag-es of faces on the web. Then we cropped them to the size 600 pixel × 600 pixels and assure all images have face elements at the same position. We categorized faces by sex, age and race.

Visual appearance age was categorized in three segments:

• Children and young people• Middle age people• Elderly people

Journal of Graphic Engineering and Design, Volume 7 (1), 2016. 7

Race was visually categorized and equally used as:

• White (Caucasian) race• Black (Negro) race• ellow (Mongolian) race

Analysis of how race influence face recognition would require further study because it has additional social and psychological aspects (Klama & Milton, 2012; Hills & Pake, 2013), so we made a general set of faces by race.

Procedure

We prepared three tests with different duration of face view. Each test contains 10 faces categorized as men-tioned above. Faces in test “1 second” were shown for one second, at test “2 second” for two seconds, and at test “4 second” faces were shown for the period of 4 seconds. Each participant completed all three tests, but we changed the order of test for three groups in the shape of Latin square shown in Table 1. Latin square ensures test to be independent from partici-pant’s concentration which is fading as time passes.

Table 1Latin square of tests order

order of teststest group a “1 second” “2 seconds” “4 seconds”test group b “2 seconds” “4 seconds” “1 second”test group c “4 seconds” “1 second” “2 seconds”

All 10 faces were shown in continuous order with 1 second pause of black screen. The purpose of that was to neutralize position of eye gaze before appearance of new face.

Experiment 1

We focused on three main face elements (Golding-er, He & Papesh, 2009; Hills & Pake, 2013): eyes, nose and mouth. To obtain required metrics of these face elements we set AOI (Area of Interest). Figure 2 shows example of our AOI on analyzed face image.

Each recording had sample percentage. Sample per-centage shows how many sent signals from eye tracker were correctly recorded. For example, if participants look away from the monitor eye tracker couldn’t record his gaze. Other causes of problem with receiving signal were glasses, women’s eye makeup, ambient light, etc.

In general, sample rate tells us the quality of recordings and according to instructions (Tobii Technology AB, 2012) we eliminated recordings under 90 % recorded samples (we treated them as bad recordings). Normally, longer face presentation had better recorded samples. This 90 % criteria was passed by 34 participants in “4 second”

test, 29 participants in “2 second” test and 27 partic-ipants in “1 second” test. In faster visual information change (shorter time of face presentation) the velocity of eye movements was higher, so, consequently, the possi-bility for eye tracker to lost track of eyes was also higher.

» Figure 2: Our AOI at face image (eyes, nose and mouth)

Within these three AOI we investigated:

• Fixation time for different tests;• How many participants returned to the AOI of

eyes after they left it (“regression level” or “rate of return”) (Goldinger, He & Papesh, 2009);

• How many participants first looked at the eyes (first fixation);

• What percentage of all time partici-pants spent looking at eyes (face ele-ment that attract most attention);

• How many participants looked at the mouth during time of face image presentation and if they did, how long they observed it;

• How many participants looked at the nose during time of face image presentation and if they did, how long they observed it;

Experiment 2

In second experiment we examined the capability of memorizing faces by test participants. The aim was to identify correct remembrance of faces in correlation with time of face presentation (Leyk et al., 2008).

Memorizing faces is very important in many fields of face identification (as an eyewitness, at border control, teachers in schools) (Divyarajsinh & Brijesh, 2013). In doing so, we tried to figure out the difference between the results of the memorizing faces, depending on the time display. Procedure was the same as in experiment 1. Participants were given presentation of ten faces

8

which were changing every one second (“1 second” test), two seconds (“2 second” test) and four seconds (“4 second” test).

» Figure 3: Sample of sheet of faces

At the end of the all presentations participants got a sheet of twenty faces. Ten were shown in the test and other ten were not in the test. Participants were re-quired to mark those which they thought that they had seen in the test.

results

Experiment 1

Table 2 shows results for AOI of eyes, mouth and nose for all investigation mentioned before.

Experiment 2

Results of successful memorization of faces are shown in Table 3. We tested 42 participants divided in three groups, who have different order of tests (“1 second”, “2 seconds”, “4 seconds”). Each group had 14 participants, thus 140 presented faces.

By changing order of tests we tried to find out whether results of recognition is worse if the test came later (per-son concentration drops with time).

test name“1 second” “2 seconds” “4 seconds”

Average fixation time (ms) 256 281 320

eyes

Time of observation (s) 0.693 1.116 2.288% of time spent observing eyes 69.3% 55.8% 57.2%Number of return to the eyes 47 189 314Max. number of return to the eyes ares 270 290 340Percentate of return (regression level) 17.4% 65.2% 92.4%First fixation made on eyes area 232 218 251All number of first fixation 270 290 340Percentage of eyes area as first fixation 85.9% 75.2% 73.8%

Mouth

Time of observation (s) 0.11 0.34 0.62% of time spent observing mouth 11.0% 17.0% 15.5%Number of visits to area mouth 106 198 276Max. number of visits to the mouth area 270 290 340Percentage of visit to the mouth area 39.3% 68.3% 81.2%

Nose

Time of observation (s) 0.15 0.25 0.43% of time spent observing nose 15.0% 12.5% 10.8%Number of visits to area nose 121 173 267Max. number of visits to the nose area 270 290 340Percentage of visit to the nose area 44.8% 59.7% 78.5%

Total time of observings all three areas (s) 0.953 1.706 3.338Percentage of observing all three areas (s) 95.3% 85.3% 83.4%

Table 2Results of observation of three main face elements (eyes, mouth and nose)

Journal of Graphic Engineering and Design, Volume 7 (1), 2016. 9

Discussion

Experiment 1

First we investigated average fixation time for differ-ent time of face presentation. It is clear that test with shorter time of face presentation had much shorter average fixation time. The differences in average fixation time in quite significant (256 ms, 281 ms and 320 ms). These results confirmed that eye quickly adapted more dynamic presentation with shorter fixation time (Dorr et al., 2010).

At the test “4 seconds” fixation time 320 ms was very consistent with normal fixation time for observing faces (Guo et al., 2006). We can say that this longer face presentation is most natural way of viewing faces. As we predicted most important face element (eyes) attracted most attention. From the time of observation area of eyes and the time of face presentation we calculated percentage of time spent at area of eyes. For test “1 second” this percentage was the highest among all tests (69,3 %), because in this test faces changed so rapidly that participants gaze rarely left the area of eyes and visited the area of other parts of the face.

At other two tests participants had enough time for ob-servation other face elements, so the percentage of time observing eyes was much smaller (55,8 % and 57,2 %).

Test “4 seconds” had slightly higher percentage than test “2 seconds”, due to the fact that participants more often returned back to the area of eyes after observing other face elements.

We also investigated “regression level” (“rate of return”) for the area of eyes. That is percentage of return to the area of eyes after leaving it to observe other face elements (Goldinger, He & Papesh, 2009). Eye tracker metrics for that information is number of visits to the AOI. If that number is more than 1, test participant had left the area and also returned back one or more times. At longer time of face presentation participants more often left the eyes area, observed other face elements and come back to the eyes.

Results showed that switching faces at one second period was much too fast for participants to perform procedure described above. So regression level was only 17,4 %. Regression level of 92,4 % at “4 seconds” has shown us that four seconds period of face presentation was long enough to observe other parts of the face (also cheeks, chin, forehead, ears and hair) and most of the test subjects have returned to the eyes area.Due to the higher regression level, the percentage of time observing eyes area in the test “4 seconds” was a little higher than in test “2 seconds”, because at test “4 seconds” partici-pants more often returned to the area of eyes.

Eyes are usually the first area for the user’s gaze. So we also explored eyes area as the first fixation. In all three tests first fixation at most of the faces were made at eyes area. The highest percentage of first fixation on eyes area was at test “1 second” (85,9 %). This can be explained with the fact that at very rapid face change participants gaze didn’t even left the area of eyes before the next face image appeared, so at the time of appearance the next face, the gaze was already at that area (despite that we used black screen for 1 second). In other two tests this percentage was lover (75,2 % and 73,8%), but still eyes definitely attracted most attentions.

Comparing results of mouth and nose showed us some interesting aspect. In “1 second” test number of visits to the target areas and time of observation was higher at nose area. One second time change was very fast for participants, so if they even left the area of eyes, their gaze usually did just a small saccade down (area of nose). In general, mouth are more attractive face elements than nose, so in our case, when participants had enough time for observation (“2 seconds” and “4 seconds” test), per-centage of visits and specially time observation of those area were in favor of mouth.

Last calculation was about all three tested areas together. As those are three most attractive face elements, partici-pants observe other parts of face only when face presen-tation was long enough. We already found out that one second was too short and sum percentage of observation of three areas are very high (95,3 %). In other words, less than 5 % of all fixations were made outside of these three main face elements.

All faces within test 140 140 140Test group A (1, 2, 4) Recognized 102 123 129

% success recognition 72.9% 87.9% 92.1%Test group B (2, 4, 1) Recognized 92 131 133

% success recognition 65.7% 93.6% 95.0%Test group C (4, 1, 2) Recognized 99 125 134

% success recognition 70.7% 89.3% 95.7%Comulative % succcess recognition 69.8% 90.2% 94.3%

Table 3Results of face recognition

10

Results for other two tests with longer time of face presentation are very similar. Manual review showed that in “2 seconds” and “4 seconds” tests majority of par-ticipants had enough time to also look other face parts of faces (cheek, forehead, ears, etc.)

Figure 4 shows samples of heat maps for one face in every test (“1 second”, “2 seconds” and “4 seconds”). In first face heat map almost all fixations were made in the area of eyes, second face heat map presents that mouth was also strongly observed and at third face heat map we can see that participants had enough time to look also at the ears and cheek. Those are most representative heat maps for different tests.

Experiment 2

As we said participants were divided in to three groups (A, B and C) and order of tests we organized in Latin square. Each set contain 14 participants; thus 140 faces were presented. Table 3 shows results for each test group. Results of correct recognition increased with lon-ger time of face presentation. But the difference between “1 second” test and “2 seconds” test was much bigger then between “2 seconds” test and “4 seconds” test. When we compared same tests in different order results also met our expectation. Recognition results for test “1 second” were the best when this test was performed first (test group A) and worst when it was performed last (test group B). Fall of concentration had some influence to the ability of memorizing. Same pattern can be seen for test “4 seconds”. Test “2 seconds” also had best recognition results when it was performed first. Cumulative results of correct recognition for tests “2 seconds” and “4 seconds” were very good (90,2 % and 94,3 %), whereas test “1 second” had much lower recognition success (69,8 %).

Conclusion

In our study we investigated how people look at the face elements in dependence from time of face pre-sentation. We confirmed that eyes are by far most attractive element. Only if participants had more

time available, their gaze also went to the mouth and nose, whose were next two face elements that attracted attention. Even in our longest test (4 sec-onds) other face element were rarely observed.

Results has shown that for memorizing faces using short term memory time of face presentation of one second was too short and two seconds was already long enough. Longer face presentation doesn’t improve results in such great level. Further research could investigate how different numbers of present-ed faces influence on face recognition success.

We can also conclude that from user experience per-spective, any presentation on internet should include person image. It is also advisable that person must face the camera.

Important factor in process of face recognition is also emotion of the faces. Authors have slightly dif-ferent set of basic emotion (happy, angry, sad, fear, surprised, annoyed, etc.). Our future research is to investigate how people see and remember faces with different emotions (face expression) and what are most attractive face elements for different emotion.

References

Buchan, J.N., Pare, M. & Munhall, K.G. (2007) Spa-tial statistics of gaze fixations during dynamic face processing. Social Neuroscience. 2 (1), 1-13. Available from: doi: 10.1080/17470910601043644 [Accessed 6th September 2014].

Cangöz, B., Altun, A., Aşkar, P., Baran, Z. & Mazman, S. G. (2013) Examining the visual screening pat-terns of emotional facial expressions with gender, age and lateralization. Journal of Eye Movement Research. 6 (4), 1-15. Available from: doi: 10.16910/jemr.6.4.3 [Accessed 7th September 2014].

Cassin, B. & Solomon, A. B. S., Melvin, L. R. (ed.) (1990) Dictionary of Eye Terminology. 2nd edition. Gainesville, Triad Pub Co.

» Figure 4: Samples of heat maps for different tests

Journal of Graphic Engineering and Design, Volume 7 (1), 2016. 11

Divyarajsinh, N. P. & Brijesh, B. M. (2013) Face Rec-ognition Methods & Applications. International Journal of Computer Technology and Applications. 4 (1), 84-86.Available from: http://www.ijcta.com/documents/volumes/vol4issue1/ijcta2013040114.pdf [Accessed 7th September 2014].

Dorr, M., Martinetz, T., Gegenfurtner, K. R. & Barth, E. (2010) Variability of eye movements when viewing dynamic natural scenes. Jour-nal of Vision. 10 (10), 1-17. Available from: doi: 10.1167/10.10.28 [Accessed 30th August 2014].

Goldinger, S. D., He, Y. & Papesh, M. H. (2009) Defi-cits in Cross-Race Face Learning: Insights from Eye Movements and Pupillometry. Journal of Experi-mental Psychology: Learning, Memory and Cogni-tion. 35 (5), 1105-1122. Available from: doi: 10.1037/a0016548 [Accessed 6th September 2014].

Guo, K., Mahmoodi, S., Robertson, R. G. & Young, M. P. (2006) Longer fixation duration while view-ing face images. Experimental Brain Research. 171 (1), 91-98. Available from: doi: 10.1007/s00221-005-0248-y [Accessed 5th September 2014].

Heisz, J. J. & Shore, D. I. (2008) More effi-cient scanning for familiar faces. Journal of Vision. 8 (1), 1-10. Available from: doi: 10.1167/8.1.9 [Accessed 25th August 2014].

Henderson, J. M., Falk, R. J., Minut, S., Dyer, F. C. & Mahadevan, S. (2001) Gaze control for face learn-ing and recognition in humans and machines. In: Shipley, T. & Kellman, P. (eds.) From fragments to objects: Segmentation and grouping in vision.Amsterdam, Elsevier Science, pp. 463-481.

Henderson, J. M., Williams, C. C. & Falk, R. J. (2005) Eye movements are functional during face learning. Memory & Cognition. 33 (1), 98-106. Available from: doi: 10.3758/BF03195300 [Accessed 23th August 2014].

Hills, P. J. & Pake, J. M. (2013) Eye-trackingthe own-race bias in face recognition: Revealing the perceptual and socio-cognitive mechanisms. Cognition. 129 (3), 586-597. Available from: doi: 10.1016/j.cogni-tion.2013.08.012 [Accessed 15th August 2014].

Hooge, I. T. C. & Erkelens, J. C. (1998) Adjustment of Fixation Duration in Visual Search. Vision Research. 38 (9), 1295-1302. Available from: doi: 10.1016/S0042-6989(97)00287-3 [Accessed 28th August 2014].

Horsley, M., Eliot, M., Knight, B. A. & Reilly, R. (eds.) (2014) Current Trends in Eye Track-ing Research. Berlin, Springer.

Klama, E. K. & Milton, F. (2012) Differences in eye movements between same and other race face recognition. In: Miyake, N., Peebles, D. &Cooper, R. P. (eds.) Building Bridges Across Cognitive Sci-ences Around the World: Proceedings of the 34th Annual Meeting of the Cognitive Science Soci-ety, 1-4 August 2012, Sapporo, Japan. Austin, Cognitive Science Society. pp. 1804-1809.

Kyrnin, J. (2014) What is Web Content? Available from: http://webdesign.about.com/od/content/qt/what-is-web-content.htm [Accessed 23thAugust 2014].

Leyk, D., Sievert, A., Heiss, A., Gorges, W., Ridder, D., Alex-ander, T., Wunderlich, M.& Rüther, T. (2008) Valida-tion of a short-term memory test for the recognition of people and faces. Ergonomics. 51 (8), 1125-1136. Available from: doi: 10.1080/00140130802094371 [Accessed 5th September 2014].

Minear, M & Park, D. C. (2004) A lifespan data-base of adult facial stimuli. Behavior Research Methods, Instruments&Computers. 36 (4), 630-633. Available from: doi: 10.3758/BF03206543 [Accessed 7th September 2014].

Nielsen, J. & Pernice, K. (2010) Eyetracking Web Usability. Berkeley, New Riders.

Patel, N. (2014) 8 Powerful Takeaways from Eye Tracking Studies. Available from: http://www.quicksprout.com/2014/04/16/8-powerful-takeaways-from-eye-tracking-studies/ [Accessed 6th September 2014].

Pollatsek, A., Rayner, K. & Balota, D. A. (1986) Inferences about eye movement control from the percep-tual span in reading. Perception & Psychophys-ics. 40 (2), 123-130. Available from: doi: 10.3758/BF03208192 [Accessed 20th August 2014].

Tobii Technology AB. (2012) User Manual Tobii Studio.Version 3.2.Danderyd, Tobii Technology AB.

Journal of Graphic Engineering and Design, Volume 7 (1), 2016. 13

Introduction

A concept of corporate social responsibility and behavior change towards sustainability and ‘greener’ print choices are motivated by different social, economic and environ-mental factors (Arikboğa, 2012; Werner, Rothenberg & Miller, 2012).Over the past few years overall office print output has decreased with a great potential to reduce it even more, which is mostly due to utilization of dif-ferent electronic devices and the emergence of various tactics which can be employed to increase the green-ness of print behaviours (Bigelow et al., 2011; Werner, Rothenberg & Miller, 2012). Some of those measures are: e-mail, projector and dual screen utilization, duplex printing, documents scanning, draft print mode utiliza-tion, single-spaced text, recycled paper, 2-up or more per page printing mode, font size below 12pt, rational typographic choices, delayed print, no banner sheet, watermark, straggler/orphan control, margins minimiza-tion, toner-save mode/decrease ink volume, no colour, print preview, no images and eco-font utilization (Bigelow et al., 2011; Werner et al., 2012). Many companies focus

on these kind of cost-effective activities, which often do not require large investments and represent very simple solutions. One of the afore mentioned measures for pro-motingsustainable office development is eco-font utiliza-tion, which had been undertaken by a Turkish company Türk Telekom (Arikboğa, 2012). Eco-fonts such as Spranq and InkSaver reduce ink consumption by placing holes or striations within the text, but user testing suggested that both were objectionable in terms of congeniality, i.e. the pleasure of reading (Bigelow et al., 2011). According to certain investigations rational typographic choices can lead to green printing improvements in terms of less ton-er consumption, where several font families were iden-tified as toner-efficient (Garamond, Times New Roman, and Century Gothic)(Bigelow et al., 2011; Mirchandani & Pinko, 2014). Besides font family selection, it is neces-sary to consider the typeface size and style for ensuring the information permanence (Možina et al., 2010). Bigelow et al. (2011) modified sans-serif open source font DejaVu and generated a new intermediate font weight called DejaVu Sans Light Thesis, using FontLab Studio software, which provided greater calculated ink

Rastko Milošević, Uroš Nedeljković, Bojan Banjanin, Dragoljub Novaković,Nemanja Kašiković

University of Novi Sad, Faculty of Technical Sciences, Department of Graphic Engineering and Design, Novi Sad, Serbia

Corresponding author:Rastko Miloševiće-mail: rastko.m@uns.ac.rs

First recieved: 21.12.2015.Accepted: 10.02.2016.

abstract

Utilization of eco-font for office printing is one of sustainable, “green” printing concepts, which besides obvious economic benefits, as a result has a certain effect on environmental sustainability as well. The fundamental problem that this practice faces is decreased quality of text printed using eco-fonts comparing to those printed with regular fonts. The aim of this research is eco-font efficiency estimation, i.e. determination of toner usage reduction level of ink jet printed documents typed with this font type, as well as estimation of the extent humans perceive differences between text printed with eco-font and the one printed by its „non-eco“ equivalent. Combining instrumental measuring method and digital image analysis, it was found that this simple principle (eco-font utilization) enables substantial toner usage reduction for an ink jet printing system, while visual test showed that visual experience of text printed using eco-font is sufficient. In addition, awareness of benefits that eco-font utilization brings, change users’ attitude towards eco-font quality.

KEy WoRDS eco-font, ink jet printing, toner consumption

the analysis of ink jet printed eco-font efficiency

Original scientific paper http://doi.org/10.24867/JGED-2016-1-013

14

savings, comparing to other tested font families (Helvet-ica, Arial, Century Gothic, Times New Roman, DejaVu: Book, Light, Extra Light) while maintaining readability.

In previous investigations researchers usually estimat-ed ink consumption as an ‘ink’ coverage percentage, i.e. the number of black pixels in the digital file used to image required text in the test target, with different soft-ware tools (Photoshop and APFill Ink & Toner Coverage Calculator)(Bigelow et al., 2011; Mirchandani & Pinko, 2014). By basing the measure of ink consumption on pixels number in a digital image and not on the actual printed ink coverage area, the problem is abstracted from particular measures of ink jet ink or laser-printer toner on various devices, which provided a relative measure that is device-independent (Bigelow et al., 2011). The measure of printed toner coverage is known as typographic tonal density or typographic tonality. As mentioned before, it refers to the relative amount of printed ink per paper area (relative blackness of gray of type on the page). By changing different type features (typeface, type size, spacing, line length...) typographic tonality can be increased or decreased (Keyes, 1993).

Factors affecting Ink Jet toner consumption

Ink jet printing as a NIP process (Non-Impact Printing) represents one of Computer to Print technologies, where impression is generated when digital printing data arrive at the printing unit, i.e. to the ink jet noz-zle system. Then, mostly liquid inks (toners) are being squirted from the nozzles directly onto the substrate, which can be made of different materials in various sizes (Petrović, Milković & Valdec, 2013). Print quality within ink jet printing technique can be quantified in various aspects. The most important ones are color reproduc-tion accuracy, color gamut range and print sharpness. In order to fulfill those print quality requirements, paper substrate materials designed for ink jet printing should therefore possess several qualities such as: sufficient hold out of ink on the material surface to provide high optical print density, rapid absorption of the ink vehicle liquid for fast drying in order to prevent feathering and bleeding, low colour-to-colour bleed (well-defined diffu-sion of the ink), low strike-through, water-fastness and light-fastness, uniform density of the paper, wicking of ink along fibers in the surface of the sheet (Durbeck, & Sherr, 1988; Svanholm, 2007; Szentgyörgyvölgyi & Bor-bély, 2015). The rate of absorption and lateral spreading of ink jet inks on paper are largely determined by the wetting, spreading, and absorption characteristics of vehicle solutions on paper. Ideally, ink vehicle should penetrate rapidly into the paper structure leaving the dyes as close as possible to the paper surface, in order to avoid smearing, running, and splatter when multiple drops arrive at the same location on the paper (Durbeck & Sherr, 1988). Toner usage of an ink jet printer, as an important factor contributing both final print quality

and overall printing costs, is affected by several factors: environmental parameters during printing process (ambi-ent temperature 23°C ± 2°C and relative humidity 50% ± 10%), printing settings (print mode, printing software, driver printer firmware, computer operating system)... (Brother Industries Ltd., 2008; Jeran, 2008). Toner usage can be defined as the weight of toner consumed during the printing of one page that has a specified coverage of text and/or graphics (standard test target is a page of text with 5% coverage of the printable area, 8” x 10” for 8.5” x 11” paper). It is calculated and reported as the number of milligrams of toner per page used (mg/page) based on the weights measured (Wyhof, 1997).

The aim of this research is determination of eco-font efficiency, i.e. estimation of the toner consumption reduction level of ink jet printed documents typed using an eco-font. Since typefaces and document layouts for office printing were not designed for the purpose of reducing toner and/or paper usage, crucial aspect, which has been covered within the paper as well, is the analysis of human visual perception of the prints, because modifications made to increase sustainabil-ity and ‘greener’ printing behaviour should not, but usually do decrease readability and congeniality.

Method

The samples inspected within the investigation were printed using Canon Pixima iP4200 ink jet office printer with PGI-5BK Black Ink, on standard office copy paper (80 g/m2 uncoated wood-free paper). Before prints were made, certain basic physical and optical properties of the selected paper were measured, which results are presented in Table 1. The paper basis weight was mea-sured according to the ISO 536 standard using ABJ-120 KERN analytical balance (International Organization for Standardization, 1995), while the paper thickness was measured according to the ISO 534 standard (Interna-tional Organization for Standardization, 1988). Surface roughness of the paper samples were measured via contact profilometry method, using Surface Roughness Tester TR 200 (5 cut-off lengths of 0,8mm; RC filter; range: ±40 µm), while mechanical characteristics of the selected paper (break force, break stress, stroke, and strain) were measured using Shimadzu EZ-LX compact table-top universal tester, according to TAPPI T 494 stan-dard (Technical Association of the Pulp and Paper Indus-try, 1996). Then, optical properties of the paper were determined, i.e. whiteness and yellowness, using spec-tro-densitometer Techkon SpectroDens (measurement geometry 0/45°, illuminant D65, standard observer 2°).

Ryman Eco font (Ryman, n.d.), as one of numerous open source, free eco-fonts, was selected for the tests. Besides original eco-font which will be tested, its modified, „non-eco“ match version was needed as well, in order to

Journal of Graphic Engineering and Design, Volume 7 (1), 2016. 15

compare the toner consumption after printing process to eco-font toner consumption. Font typefaces modifica-tions referred to enclosing of open typeface structures in order to generate font version with “filled” typeface appearance, but other typeface characteristics remained the same as the original one (Figure 1). Mentioned mod-ifications were done using FontLab Studio 5 software. Then, two test charts were designed, both consisting of the same text elements in different sizes from 8pt to 16pt, but in the first one text was typed using eco-font, and the other one consisted of modified, non-eco font.

Table 1 Physical and optical characteristics of the tested paper

Basis weight [g/m2] 80Thickness [mm] 0,1067Specific volume [cm3/g] 1,33

Surface roughness (Ra) [μm]MD CD

2,652 2,937Break force [N] 120,98 61,36Break stress [N/mm] 4,84 2,45Stroke [mm] 2,85 6,58

Strain [%] 1,90 4,38

Whiteness (WCIE) 124,00Yellowness (Y1925) -19,61

There were the total of 40 printed A4 format (210 x 297mm) paper samples, 20 printed using eco-font and the other half of the samples printed with modified, non-eco font. Initial toner consumption calculation was based on the mass measurements of the sam-ples before and after printing process using men-tioned analytical balance, where each blank paper sample was firstly measured, then printed, and right after printing process it was measured again, which was repeated for each paper sheet (under relative air humidity of 52% and air temperature of 23°C). When printing and measuring process of the samples’ masses were done, samples were digitalized using flatbed scanner CanoScan 5600F, on 1200 spi (using white backing, without using any correction filter).

» Figure 1: Principle schema of close-range photogrammetric process

Scanned images were (TIFF and PDF file formats) need-ed for processing stage using different softwares (APFill Ink & Toner Coverage Calculator and Photoshop CS6), in order to estimate toner consumption, but now based

on the printed toner coverage of the paper sheet area. For visual comparison between text segments printed using both fonts, digital portable microscope ViTiny VT300 was used to generate enlarged images. Also, a visual method was employed to investigate respondents’ capability to distinguish the visual difference between texts printed using two different fonts. Visual test exam-ined respondents’ capability to distinguish the visual difference between text paragraphs printed with the eco (Ec) and the non-eco font (NEc). It included 12 neutral grey cardboard cards (240 x 100 mm), consisted of the two same dummy text paragraphs (black text on white paper) with the same point size and typographic features (spacing, leading, number of words). Dummy text was used to direct respondents’ attention only on visual dif-ferences between text printed using two different fonts. Fonts in four point sizes were used: 8pt, 9pt, 10pt and 11pt. For each point size, three different cardboards were used, each containing printed and cut paper samples with two paragraphs that have one of the following font combinations: Ec-Ec, NEc-NEc and Ec-NEc or NEc-Ec.

Respondents were instructed to observe each of these 12 cards separately and to state if they see any visual differ-ence between two paragraphs with the same typograph-ic characteristics (spacing, leading, number of words). They are instructed not to read paragraphs because of the dummy (pseudo) words, in order to be focused only on visual differences between the two paragraphs on each presented card. If participants notice any difference between the two paragraphs, they were instructed to state which one they prefer as more visually pleasing.

At the end of the visual test, respondent’s were informed about characteristics of two fonts, and were asked then whether they would choose to print their documents using eco-font if they knew that this font would reduce their toner consumption. In visual test participated thirty persons (thirteen males and seventeen females), third grade students of Graphic engineering and design at the Faculty of Technical Sciences in Novi Sad, aged from 21 to 25 years old with normal to corrected vision, who have certain experience in visual arts and typography. Visual test was conducted in the darkened room using Agile Radiant Controlled Light 5 color viewing cabinet (CIE Standard Illuminant D65, a normal reading distance of 40 cm) (Faye, 1984; Legge et al., 1985a; Legge et al., 1985b).

results

This part of the paper will present conducted analysis and obtained results of toner usage reduction which can be achieved when printing using eco-font, which was estimated via digital image analysis and using instru-mental method. After those objective examinations, a visual test of the printed samples was performed.

16

Ink consumption analysis

Obtained ink coverage percentages of printed samples (scans) and the two test chart PDF files (digital origi-nals with the eco and the non-eco font) used to print the samples, are presented in Table 2. Two different softwares, APFill Ink & Toner Coverage Calculator and Photoshop CS6, were used for the ink coverage percent-age estimation. Generated APFill software results show that test chart (digital original) consisting of non-eco font has almost 33% higher ink coverage (8,79%) com-paring to the other test chart (digital original) consist-ing of eco-font (5,89%). Obtained ink coverage results using Photoshop image processing, show 45,58%higher ink coverage of test chart consisting of non-eco font (14,13%) comparing to test chart with eco-font (7,69%).

Regarding printed samples, the ones printed using non-eco font show higher toner coverage, i.e. higher toner consumption for 6,94% (APFill, 13,86%) and 7,23% (Photoshop, 15,95%), comparing to the samples printed with eco-font (APFill 12,96% and Photoshop 13,83%), which is considerably lower in comparison to previ-ously determined test chart ink coverage results. This disproportion of obtained ink coverage ratios between test charts and printed samples is the result of printed ink jet toner spreading on the paper surface, which occurs due to absorbing and fibrous nature of the paper substrate material and rheological ink characteristics.

Table 2Ink coverage percentages

Eco-font Non-Eco fontTest chart - APFill 5,89 % 8,79 %Test chart - Photoshop 7,69 % 14,13 %Samples - APFill 12,96 % 13,86 %

Samples - Photoshop 13,83 % 15,95 %

In Table 3 are presented obtained average and over-all mass values of the paper samples before and after printing process using two different font types (eco and non-eco), toner mass values, as well as per-centage of savings when printing using eco-font.

Table 3Мasses of the samples before and after printing process

Eco-font Non-Eco fontMass values average overall average overallPaper [g] 4,9565 99,1290 4,9440 98,8796Paper+toner [g] 4,9961 99,9221 5,0091 100,1824

Toner [mg/page] 39,7 793,1 65,1 1302,8

Saving [%] 39,02 39,12 - -

Instrumental measuring method, based on samples’ mass change after the printing process, enabled

more reliable results of ink jet toner consump-tion comparing to digital image analysis method. Obtained results indicate that by using the eco-font, the toner consumption is reduced by 39,02%, for this particular text included in the test charts.

Visual analysis

In Figure 2 are presented results of the visual test, i.e. respondents’ capability to distinguish the visual differ-ences between text paragraphs printed with the eco (Ec) and the non-eco font (NEc). Results of the analysis are presented via calculated paired-comparison index values (PC-index). Thurnstone’s law of comparative judgment with the statistical assumptions of a case 5 solutionwas applied to the obtaine data (respondents’ grading of the text paragraphs printed using eco and non-eco-font) in order to calculate an interval scale of the samples (Engeldrum, 2000). The PC-index scales the samples from 0 to 200, where high values corre-spond to the samples perceived as having a good text quality, while low values correspond to samples per-ceived as having poorer text quality. The PC-index was calculated according to the following equation (1):

where n is the number of observers, m is the num-ber of samples and vi is the value given by observer (value 2 for visually more pleasing font type, value 1 if observer do not notice any visual difference between fonts, and value 0 for visually less pleasing font type).

» Figure 2: Calculated PC-index for eco-font (Ec) and non-eco font (NEc) for four different font sizes

1)

Journal of Graphic Engineering and Design, Volume 7 (1), 2016. 17

Results, i.e. calculated PC-index values, show that font size increase leads to greater perceptibility of differenc-es between two printed font types, thus perceived text quality of the eco-font (Ec) decreases, which implies that perceived quality of text printed using non-eco font (NEc) increases (for all tested font sizes). This trend of the results was expected, because by increasing font size text becomes physically bigger, thus more eco-font features and imperfections (open typeface regions uncovered with ink that exist due to insufficient ink bleeding on the paper surface - bigger the font size, bigger the letter stroke gaps) are revealed, as well as differences between two font types becomes more noticeable (Figure 3).

» Figure 3: Visual comparison of different point size text segments printed using eco-font (upper row)and non-eco font (lower row)

Conclusion

Presented investigation of ink jet toner usage reduction of printed documents typed with an eco-font, showed that it is possible to save the substantial amount of ton-er using this font type. Both image analysis and mass measurements indicated certain toner consumption reduction, where the later method results showed approximately 39% ink jet toner reduction with the eco-font utilization. Since typefaces for office printing were not designed with the intention to reduce toner usage, crucial aspect is that modifications made to increase sustainability do not, or as little as possible decrease readability and congeniality, if printed communication is to remain effective, because, above all, the main pur-pose of printing is communication (Bigelow et al., 2011).Regarding visual test, almost all participants noticed visual differences between the eco-font and the non-eco font, preferring non-eco font as visually more pleasing. Even though analysed font characteristic are on the side of non-eco font, 53,33% of the respondents (16 out of 30) indicated that they would use eco-font type if they knew that this font will save the toner. This indicates that respondents are aware of visual differences between two fonts, but when it comes to sustainable and eco-nomic benefits that users can experience using eco-font, visual differences and reading difficulties becomes less important. Another explanation of the obtained results could lie in the fact that all the participants were trained in visual communication and typography (third grade

students of Graphic engineering and design), so they were prone to spotting visual differences between two font types better, thus testing participants who have no prior knowledge and experience in visual communica-tion and typography may enable more reliable results.

Further research should be pointed towards inclu-sion of more printing substrate types, different ink jet printing systems, as well as in the inclusion of electro-photographic printing systems, and its toner consump-tion estimation in eco-font printing. Also, it would be significant to extend present visual tests to legibility and readability determination of printed eco-font.

Acknowledgements

This research was supported by the Serbian Ministry of Science and Technological Development, Grant No.:35027 “The development of software model for improvement of knowledge and production in the graphic arts industry”.

References

Arikboğa, F. Ş. (2012) From Environmental Respon-sibilities to Green Management: Case of Turkey and a Comparative Analysis. İstanbul Üniversitesi Siyasal Bilgiler Fakültesi Dergisi. 0 (47), 31-58. Available from: http://dergipark.gov.tr/iusiyasal/issue/629/6392 [Accessed 3rd February 2016].

Bigelow, C., Dowd-Hinkle, D. J., Horowitz, D. & Voorhees, G. (2011) Sustainable by design: An investigation into ecologically friendly typography. Rochester, Printing Industry Center at RIT. Available from: http://schol-arworks.rit.edu/cgi/viewcontent.cgi?article=1091&-context=books [Accessed 20th December 2015].

Brother Industries Ltd. (2008) Brother Genuine Mono-chrome Toner Cartridge Testing Method for Determin-ing State Page Yield Based on ISO-IEC19752 Standard. Available from: http://download.brother.com/pub/com/iso_yield/mono_19752/declared_page_yield_iso19752_en.pdf [Accessed 20th December 2015].

Durbeck, R. C. & Sherr, S. (eds.) (1988) Digital Output Hardcopy Devices. San Diego, Academic Press, Inc.

Engeldrum, P. (2000) Psychometric Scaling - A Toolkit for Imaging System Development. Winchester, Imco-tec Press.

Faye, E. E. (1984) Clinical Low Vision. 2nd edition. Boston, Little, Brown & Co.

International Organization for Standardization. (1988) ISO 534:1988. Paper and board: Determination of thickness and apparent bulk density or apparent sheet density. Geneva, International Organization for Standardization.

International Organization for Standardization. (1995) ISO 536:1995. Paper and board: Determination of grammage. Geneva, International Organization for Standardization.

Jeran, P. (2008) Supply standards: Past and future.

18

Genève, International Organization for Standardization. Available from: http://www.iso.org/iso/livelinkgetfile-isocs%3FnodeId%3D15050486 [Accessed 15th November 2015].

Keyes, E. (1993) Typography, Color, and Informa-tion Structure. Technical Communication. 40 (4), 638-654. Available from: http://www.jstor.org/stable/43090213 [Accessed 3rd February 2016].

Legge, G. E., Pelli, D. G., Rubin, G. S. & Schle-ske, M. M. (1985a) Psychophysics of Read-ing–I. Normal Vision. Vision Research. 25 (2), 239–252. Available from: doi: 10.1016/0042-6989(85)90117-8 [Accessed 4th February 2016].

Legge, G. E., Rubin, G. S., Pelli, D. G. & Schle-ske, M. M. (1985b) Psychophysics of Read-ing–II. Low vision. Vision Research. 25 (2), 253–265. Available from: doi: 10.1016/0042-6989(85)90118-X [Accessed 4th February 2016].

Mirchandani, S. & Pinko, P. (2014) A Simple Print-ing Solution to Aid Deficit Reduction. Jour-nal of Emerging Investigators. 1-5. Available from: https://www.emerginginvestigators.org/articles/a-simple-printing-solution-to-aid-defi-cit-reduction [Accessed 3rd February 2016].

Možina, K., Medved, T., Rat, B.& Bračko, S. (2010)Influence of Light on Typographic and Colorimet-ric Properties of Ink Jet Prints. Journal of Imaging Science and Technology. 54 (6), 60403-1-60403-8(8). Available from: doi: 10.2352/J.ImagingSci.Tech-nol.2010.54.6.060403 [Accessed 10th January 2016].

Petrović, V., Milković, M. & Valdec, D. (2013) Kom-paracija karakteristika ink-jet otisaka dobivenih vodenim, solventnim i UV bojilima. Tehnički glas-nik. 7 (2), 191-197. Available from: http://hrcak.

srce.hr/index.php?show=clanak&id_clanak_jezik=155353 [Accessed 10th January 2016].

Ryman (no date) Ryman Eco (Version 1) [Font Soft-ware] Ryman. Available from: http://www.rymaneco.co.uk/ [Accessed 8th January 2016].

Svanholm, E. (2007) Printability and Ink-Coating Interac-tions in Inkjet Printing. PhD thesis. Karlstad University.

Szentgyörgyvölgyi, R.& Borbély, Á. (2015) Weather resistance of inkjet prints on plastic substrates. Journal of Graphic Engineering and Design. 6 (1), 11-14. Available from: http://www.grid.uns.ac.rs/jged/download.php?fid=151 [Accessed 10th January 2016].

Technical Association of the Pulp and Paper Indus-try (1996) T 494 om-96. Tensile properties of paper and paperboard (using constant rate of elongation apparatus). Atlanta. Technical Asso-ciation of the Pulp and Paper Industry.

Werner, K., Rothenberg, S. & Miller, J. (2012) Drivers of greener print choices: an exploratory study of print consumer environmentalism. Rochester, Printing Industry Center at RIT. Available from: http://schol-arworks.rit.edu/cgi/viewcontent.cgi?article=1101&-context=books [Accessed 22nd December 2015].

Wyhof, J. (1997) Page Yield of Printer Cartridges. System Support Series 112. Stanford, Static Con-trol Components. Available from: http://www.arbikas.com/view/storage/articles/SSS112TECH-PageYield.pdf [Accessed 12th November 2015].

Journal of Graphic Engineering and Design, Volume 7 (1), 2016. 19

Introduction

3D printing is a process of creation three-dimensional solid objects from adequately prepared files, in which design is diagonally shaped. For this printing technique a term “Rapid Prototyping” is often used, thus it desig-nates technologies, which deposit materials for 3D model construction in several layers. Different shapes and forms of objects are created with additive process in which sequential layers of diverse material are added one on the top of another. The process starts with the produc-tion of virtual object in a 3D modelling software or by using 3D scanner, which creates an object copy (3DPrint-ing.com, 2015; Muck & Križanovskij, 2015; Thomas & Claypole, 2016; ). 3D printers use different technologies, which significantly differ in the method of applying layers one on another. American Society for Testing and Mate-rials (ASTM) has defined seven different technologies of additive procedures, which enable creation of 3D objects:

1. Vat photopolymerization; 3D printer uses photopolymer resin, which polymerizes under UV light. The most common type of this technology is stereolithography.

2. Material jetting; the material is applied in small droplets from a small nozzle, layer by layer. Created 3D object is hardened (polymerized) by UV light.

3. Binder jetting; basic powder material and liquid binder are used in this technology. The powder material is spread in uniform layer, while the binder is added on the top through the nozzle following with another layer of powder, layer of binder etc. repeating until 3D object is formed.

4. Material extrusion; the most commonly used process, which presents targeted loading – Fused deposition modeling, where 3D objects are formed by depositing heated material layer by layer (material is heated inside the nozzle). Material hardens almost immediately after leaving the nozzle.

5. Powder bed fusion; most often includes so-called Selective laser sintering, where a laser is used as a power source for binding i.e. sintering powdered materials of plastic, metal, ceramics and glass, into 3D form.

6. Sheet lamination; this technology can be divided into two processes –Ultrasonic Additive Manufacturing (UAM) and Laminated Object Manufacturing (LOM). UAM mainly uses sheet laminators for the production of metal multilayer

Raša Urbas, Matej Pivar, Urška Stankovič Elesini University of Ljubljana, Faculty of Natural Sciences and Engineering,Snežniška 5, 1000 Ljubljana, Slovenia

Corresponding author: Raša Urbase-mail: rasa.urbas@ntf.uni-lj.si

First recieved: 22.04.2016.Accepted: 20.05.2016.

abstract

The aim of the research was to produce tactile floor plans for blind and visually impaired people for the use in the museum. For the production of tactile floor plans 3D printing technique was selected among three different techniques. 3D prints were made of white and colored ABS polymer materials. Development of different elements of tactile floor plans, as well as the problems and the solutions during 3D printing, are described in the paper.

KEy WoRDStactile floor plan, blind and visually impaired, 3D printing, maps, floor plan, tactile museum

development of tactile floor plan for the blind and the visually impaired by 3d printing technique

Original scientific paper http://doi.org/10.24867/JGED-2016-1-019

20

sheet constructions, which are welded together ultrasonically by using outer force with additional Computer Numerical Control (CNC) for shaping the final object, while LOM uses similar technology, but with paper or polymers, linked together with different adhesives. Later technique uses crosshatch for easier removal of excessive material after object construction.

7. Direct energy deposition; the technology is most often used in high-tech metal industry. Metal material, provided in powder form or wire, is deposited from the nozzle onto the surface of the object. On its way, the material is melted using laser, electron beam or plasma arc, and layer by layer constructed into an object (3DPrinting.com, 2015; Thomas & Claypole, 2016).

3D technology is suitable for printing different 3D objects, but it can also be successfully used for printing communicational tools for blind and visually impaired e.g. Braille, 3D maps, etc. Until recently, Braille was mainly printed with embossing, less often by screen printing and other printing techniques. Nowadays, 3D printing is becoming more and more popular, not only for printing Braille but also for printing tactile objects, maps, floor plans etc. (Jaquiss, 2011; Butler Millsaps, 2014). However, it is important to emphasize that the costs of manufacturing 3D tactile elements are still relatively high compared to the conventional printing procedures, due to which this new technology hasn’t fully asserted in the area of printing for the blind and the visually impaired.

Production of tactile floor plans require a lot of ingenuity. They are usually made of different materials like wood, metal, plastics etc., by different techniques, which are expensive and time consuming and have different advan-tages as well as disadvantages. Tactile floor plans, for example, can be made by the use of the microencapsu-lated paper, but their mechanical resistance and durabil-ity are very poor, therefore their exploitation time of use is limited. Thus, the aim of our research was to produce

adequate tactile recognizable and above all resistant and durable tactile floor plans by using 3D printing.

Tactile floor plans made in our research, were developed for the Technical Museum of Slovenia, Museum of Post and Telecommunication (Polhov Gradec, Slovenia). Our first idea was to print tactile floor plans with screen printing technique, where we wanted to use special 3D expandable printing inks, however the idea was abandoned due to the questionable durability of prints. In view of the above, we were trying to find a suitable technique, which would enable production/printing of tactile floor plans with sufficient durability, precision and adequate height. Therefore, 3D printing was selected, because we’ve assumed that with this technique proper results could be achieved.

Materials and methods

Materials

In our research three differently colored ABS poly-mer material has been used for printing tactile floor plans – white, red and blue (3D Systems, USA), which properties are presented in Table 1.

Table 1Properties of 3D printing polymer material

Property Value

Composition acrylonitrile butadiene

styrene (ABS)

Filament thickness [mm] 1.75

Color [-] white, blue, red

Softening temperature [°C] 105

Degradation temperature [°C] > 300

Density [g/cm3] 1.05

a) b)

» Figure 1: 3D printer CubePro Duo (a) and presentation of printing by extruding the polymer from a nozzle (b)

Journal of Graphic Engineering and Design, Volume 7 (1), 2016. 21

Printing procedure

In the research 3D printer CubePro Duo (3D systems, USA) was used for printing tactile floor plans (Figure 1a). It enables the use of different polymer materials such as acrylonitrile butadiene styrene (ABS), polylactide (PLA), nylon and water-soluble thermoplastic supporting material (ABS). CubePro Duo printer prints objects by the extrusion of material from the cartridge through the printing nozzle. Material is extruded through the printing head in a thin filament, 0.4 mm in diameter. Movement of the printing head is coordinated by the printing plate, which gradually lowers its position after each printed layer. New layer is printed onto previous, therefore an object is printed from the bottom up (Fig-ure 1b) (3D Graphic software, 2012; 3D Systems, 2015).

Printer enables two-color printing. When printing monochrome objects, maximal dimensions of 24.29 cm × 23 cm × 27.04 cm be printed, while when printing in two-color those dimensions are reduced. Individually printed layers can be produced in three variables: 70, 200 or 300 μm.

Graphic layouts for tactile floor plans were designed in Blender (Blender Institute, Netherlands), an open source software. Printing conditions are presented in Table 2.

Methods

Samples of printed tactile floor plans were analyzed by scanning electron microscope (SEM, JSM 6060 LV, Jeol) and digital microscope Dino-Lite Pro AM-413T (AnMo Electronics Corp., USA).

Results and discussion

Production of tactile floor plans started with the mini-mization of information to basic e.g. most important. With the help of floor plan blueprints of museums first and second floor, first suggestions of graphic layouts were designed in Blender (Figure 2). On designed graphic layouts of tactile floor plans walls, staircase, doors and a point, which indicates the location of the visitor, were marked. In accordance with this, the

Property Value

Nozzle temperature [°C] 260

Chamber temperature, providing constant printing conditions [°C]

60

Nozzle diameter [mm] 0.4

Filament diameter [mm] 1.75

Model preparation:

- resolution 200 μm (5 printed/built layers in 1 mm2)- filler type: »strong« (50% fill)- filler shape: diamante- without support materials

Table 2Printing conditions on 3D printer Cube Pro Duo

» Figure 2: First designed suggestions of tactile floor plan of (a) the first and (b) the second floor in the museum

a) b)

22

legend with elements marking museum’s path direction, staircase, thresholds and point of visitor’s location, was added. Initial graphic layouts of tactile floor plans were designed according to different types of tactile maps found in the literature (Rener, 1992; Schuffelen, 2002; Kaneko & Oouchi, 2010; Braille Authority of North America (BANA) & Canadian Braille Authority (CBA), 2011; 3ders.org, 2014; Butler Millsaps, 2015).

Initial graphic layouts of tactile floor plans were later slightly modified as follows:

• initially designed tactile staircase (Figure 3a) was through an intermediate stage (Figure 3b) redesigned into a tactile element, con-sisted of six parallel thicker lines (Figure 3c);

• we intended to print the names of all exhibi-tion rooms in Braille and Latin alphabet, so that the tactile floor plans would be usable/suitable also for normally sighted people;

• the path, from the staircase to the first exhibition room, was marked with small isosceles triangles;

• the same, but smaller isosceles triangles were used for marking the space between the doors, where hazardous, raised thresholds, visible to normally sighted, are located (Figure 3c).

On the basis of the first graphic layout designs, the fol-lowing preliminary prints with 3D printer were printed: a) prints of Braille;b) prints of Braille and Latin andc) prints of walls, defining the exhibition area.

Ad a.) Printing Braille on the 3D printer didn’t cause any major problems. Several printing attempts were performed mainly because we wanted to print Braille dot with parameters, which would meet the standards of the height (0.5 mm) and the diameter of Braille dot (1.5 mm), as well as the distance between the center of two Braille dots (2.5 mm) (Deutsches Institut für Normung e. V., 2007; Österreichisches Normungsinsti-tut, 2008; Fajdetić, 2012; International Standardization Organization, 2013). The appearance and the cross section of preliminary printed dots are shown in Figure 4. The Braille dots were composed of four layers, had adequate height (0.6 mm) and diameter (1.6 mm – 1.8 mm) (Figure 4c), however they had very sharp and disturbing edges (Figure 4a). Dots were thus appro-priately treated (polished), thus the edges became smoother and more pleasant to touch (Figure 4b).

Ad b.) The initial idea was to print tactile floor plans, which would be suitable for blind and visually impaired, as well as for normally sighted, but during printing we came across some serious problems. 3D printing technique is still in its developing phase and therefore, despite its great advantages, it also has some limitations. Our goal was to print Latin letters in red on the white base of tactile floor plans, so the letters would be clearly visible for visually impaired and normally sighted. But as the 3D technology uses the method of rending, printing (e.g. layering) red Latin letters would consequently cause them to stand out from the liner (base). Such prints wouldn’t be appropriate for blind and visually impaired, since the letters would actually be an additional dis-

» Figure 3: Faze development of initially designed (a), an intermediate stage (b) and final design of tactile staircase element (c)

a)

a)

b)

b)

c)

c)

» Figure 4: Untreated Braille dot produced by 3D printer (a), treated Braille dot (b) and its cross-section (c) (SEM; 22× and 20× magnification)

Journal of Graphic Engineering and Design, Volume 7 (1), 2016. 23

turbing factor in the tactile perception. To avoid this unwanted effect, we have tried to “engrave” the red Latin letters into the white base, so they would be in the same level as the liner. The attempt failed. Namely, to achieve the desired result, it was necessary to redesign the graphic layout of the 3D model in Blender. For this operation the transformation tool Boolean Modifier was selected. Tool creates a single compound object out of two mash objects, using its own algorithm of disjunction (intersection), by which empty spaces, in the shape of individual letters, were formed (Figure 5a above). Unfor-tunately, newly designed 3D mash model of text (e.g. letters) became extremely complex (Figure 5a below). The excessive amount of information caused inadequate performance of the print nozzle, resulting in emerging errors in the shape of unevenly deformed surfaces of prints (Figure 5b). This problem could be eliminated by repairing the entire 3D mash (every surface, every edge and every vertex) separately, adjusting each element (letter) to the print nozzle, which proved to be too com-plicated. In addition, another problem occurred, arising from the identical size of individual elements (e.g. letters) taken from the liner and ones in the liner. Because of the identical size the same position (wall) was printed twice – once printing the red letters and once printing the white liner. Moreover, the extrusion line slightly expanded and became more warp. Altogether caused unevenness and irregularities of printed surface (Figure 5b). Hence, in modeling of each element (e.g. letter) an addition of expanding factor needed to be considered (in our case 0.4 mm). This modification of the 3D mash revealed to be time consuming, due to the fact that the elements (e.g. letters) could not be symmetrically

increased or decreased. Similar occurred with printing numbers, therefore hereafter we have focused only on Braille, which was printed without any major problems.

Ad c.) Printing of walls, which define the individual exhi-bition area in the museum, and the Braille dots caused no problems. It was established after printing, that the wall height was a bit too high (4 mm). Therefore, in the next printing attempt the wall height was reduced to 2 mm, which was still sufficient for tactile recognition. On the bases of the preliminary printing attempts and the first printed graphic layouts of 3D floor plans we’ve decided:

• to include the names of exhibition area only in Braille and to exclude ones in Latin;

• that the tactile staircase element is adequate, but it extends too far into the hall area (Figure 2b), so we decided to reduce the number of tactile staircase element lines from six to four;

• that by reducing the tactile staircase ele-ment length more space was created on the floor plan, so it was possible to circularly allocate arrow shaped isosceles triangles, indicating museums path from the stair-case to the exhibition area (Figure 6);

• to add the legend of individual elements repre-senting: staircase (three thicker parallel lines), museums path and raised door thresholds between separate museum exhibition areas (isosceles triangle), and the dot (small round circle) denoting the visitor’s location inside the museum.

a) b)

» Figure 5: Complex mesh of the 3D model of text and (a) unsuccessful attempt of simultaneous 3D printing of Braille and Latin letters (b)

24

Based on the findings, we have designed new graphic layout of tactile floor plans in Blender (Figure 6).

In accordance with designed graphic layouts we’ve tried to produce a tactile floor plan in red and white combination. On the 3D printer simultaneously white liner, including Braille dots, and red elements were printed. Two printing heads have been used for printing white and red ABS polymer. When the printer achieved the height, on which red elements started to print, firstly two red layers were alternately printed, then two white layers, again two red and so on, until the adequate height of 3D design was achieved. Due to the size of the tactile floor plan and the dimension limitations of the 3D printer, designed tactile floor plan was divided into two parts, which were printed separately and then later joined on the basis of Plexiglas. Prototype of described tactile floor plan is presented on Figure 7.

Prototype of tactile floor plan was successfully produced, but it still had some irregularities:

• orientation of the exhibition area was inappropriate according to the posi-tion of the tactile floor plan;

• four lines, presenting staircase, were still reaching too far into the hall area;

• there was a wall missing in designed graphic layout;

• graphic layout of some doors was not correct according to the actual positioning inside exhibi-tion area, etc.

Listed irregularities were corrected and redrawn in Blender, where a new graphic layout was designed. This new tactile floor plan was firstly printed on the microencapsulated paper and tested among children at the Institute for blind and visually impaired youth (Ljubljana, Slovenia).

With the help of practice tests we have established that:

• arrow shaped isosceles triangles, which were marking the museums path from the staircase to the first exhibition area, were completely unusable (Figure 8a), namely blind children were too engaged defining where isosceles triangles were pointing, instead of using them as a direction markers. Therefore it was decided to change isosceles triangle shaped arrows into equilateral triangles (e.g. empty arrow heads), as shown on Figure 8b.

• the Braille cell (e.g. letter) was too large, and the children had difficulties reading the content; it was decided that the Braille cell needed to be slightly reduced, in particular the distance between individual letters. Later, it was established that the error occurred during the conversion of text from the Latin alphabet into the Braille typeface, which was performed in Word (Microsoft). Namely, this software conversion doesn’t fully consider the definitions of Braille standard, therefore individual Braille cells were too remote from each other;

• because of the larger Braille cell, the names of exhibition areas were considerably clustered between the walls of the tactile floor plan, with the first and last letters almost touching the walls. Due to mentioned, children were almost unable to detect the beginning and the ending of the exhibition area name;

• the walls of exhibition areas, other tactile elements and Braille were adequate in height, and as such distinguishable from the liner.

Based on the findings once again we designed new, revised graphic layout of 3D floor plan (Figure 8), printed them with microencapsulated paper and again verified with blind and visually impaired children.

Research has shown that:

• regardless to the differently designed shape of arrows designating the museum’s path, they continued to dismay the children. Therefore, it was decided that the museum’s path will not be designated and only arrows, marking thresholds between the doors, were left;

• the children were able to more easily read the narrowed and slightly reduced Braille;

» Figure 7: Prototype of tactile floor plan

» Figure 6: Redesigned 3D floor plan

Journal of Graphic Engineering and Design, Volume 7 (1), 2016. 25

• the children liked Braille printed with 3D printer, which was, due to the sharp edges, appropriately treated (polished);

• the children (and their teachers) were impressed by the idea that the walls on 3D floor plans were slightly higher than other elements.

Responses of the children led us very close to the final design of tactile floor plans. When all adjustments and corrections were made, tactile floor plans were printed, this time in blue and white combination, since blue is one of the museums colors. Printing was performed with one printing head, where first white liner with Braille was printed, and secondly blue walls and other elements were separately printed and later glued on the white liner. This method of producing tactile floor plans was used because of the difficulties, which occurred during printing i.e. fluctuation of temperature. Change in the temperature caused high shrinkage and

twisting of the ABS polymer, for which simultaneous printing of both blue and white ABS polymer was impracticable. Printed and glued tactile floor plans (Figure 9a) were fixed on the basis of Plexiglas. Finally, completed tactile floor plans were mounted on the walls in an appropriate hands height, so the blind and the visually impaired could touch them (Figure 9b).

Conclusions

3D printing technique is absolutely appropriate for printing tactile floor plans. With the exception of simultaneous printing of Braille and Latin alphabet, which we weren’t able to appropriately develop, we have successfully designed and printed all other pre-ferred elements, both in terms of the shape as well as the height. Prints made on 3D printer were slightly rougher, but with the appropriate treatment of the

a) b)

a) b)

» Figure 9: Detail of the final tactile floor plan (a) and the final tactile floor plan of toilet facilities mounted on the wall (b)

» Figure 8: Comparison of two newly designed tactile floor plans with differently shaped arrows designating the museum’s path (a and b)

printed designs we have managed to smooth the edges in a manner so they were pleasant to touch and completely unobtrusive even for children.

When printing tactile floor plans attention should be made to specific details, which normally sighted otherwise see as “perfect” solutions, but later turn out to be completely useless for the blind and the visually impaired (e.g. marked paths). Therefore, when designing of tactile maps for the blind and the visually impaired, it is absolutely necessary to coop-erate with them during the prototype designing.

References

3ders.org. (2014) Japan develops software for print-ing 3D maps for the blind. Available from: http://www.3ders.org/articles/20140926-japan-de-velops-software-for-printing-3d-maps-for-the-blind.html [Accessed 3rd January 2015].

3D Graphic software. (2012) What’s the difference between 2D and 3D anyway? 3D Graphic soft-ware. Available from: http://3dgraphicssoftware.blogspot.com/2012/06/what-difference-between-2d-and-3d.html [Accessed 5th February 2015].

3DPrinting.com. (2015) What is 3D printing? Available from: http://3dprinting.com/what-is-3d-printing/ [Accessed 28th March 2016].

3D Systems. (2015) CubePro 3D Printing. Real. Pro. 3D Systems. Available from: http://cubify.com/cubepro [Accessed 28th November 2015].

Braille Authority of North America (BANA) & Canadian Braille Authority (CBA) (2011) Guidelines and Standards for Tactile Graphics, 2010.Braille Authority of North America, Available from: http://www.brailleauthority.org/tg/index.html [Accessed 28th November 2015].

Butler Millsaps, B. (2014) 3D maps created by IDeA center use multi-sensory installations to delight and direct everyone. 3DR Holdings. Available from: http://3dprint.com/26307/3d-maps-idea-center/ [Accessed 8th August 2015].

Butler Millsaps, B. (2015) Students created 3D printed tactile map of campus for visu-ally impaired. 3DR Holdings. Available from: http://3dprint.com/63896/3d-printed-tactile-campus-map/ [Accessed 19th January 2015].

Deutsches Institut für Normung e. V. (2007) DIN 32976. Braille – requirements and dimensions, Berlin, Deutsches Institut für Normung e. V.

Fajdetić, A. (2012) Standardisation of Braille in the EU and other European Countries. In: Schwan, J. & Kahlisch, T. (eds.) Proceedings of World Congress Braille21 Innovations in Braille in the 21st Century, 27-30 September 2011, Leipzig, Germany. Leipzig, German Central Library for the Blind in Leipzig. pp. 287-294.

Jaquiss, R. S. (2011) An introduction to tac-tile graphics. Journal of Blindness Innova-tion and Research. 1 (2). Available from: doi: 10.5241/2F1-6 [Accessed 9th August 2015].

International Standardization Organization. (2013) ISO 17351:2013. Packaging – Braille on pack-aging for medicinal products. Geneva, Inter-national Standardization Organization.

Kaneko, T. & Oouchi, S. (2010) Tactile Graphics in Braille Textbooks: The Development of a Tac-tile Graphics Creation Manual. NISE Buletin. 10, 13-28. Available from: http://www.nise.go.jp/kenshuka/josa/kankobutsu/pub_a/nise_a-10/nise_a-10_2.pdf [Accessed 30th November 2015].

Muck, T. & Križanovskij, I. (2015) 3D – Printing [3D – Tisk]. Ljubljana, Založba Pasadena.

Österreichisches Normungsinstitut. (2008) ÖNR 2915753:2008. Packaging – Package Leaflets For Medicinal Products – Braille And Other Formats For Visually Impaired People. Wien, Österreichisches Normungsinstitut.

Rener, R. (1992) Tactile charts and diagrams [Taktilne karte in diagram]. MSc thesis. University of Ljubljana.

Schuffelen, M. (2002) On Editing Graphics for the Blind. The Hague, Netherlands Library for Audio Books and Braille. Available from: http://piaf-tactile.com/docs/Tactile_Graph-ics_Manual.pdf [Accessed 9th August 2015].

Thomas, D. J. & Claypole, T. C. (2016) 3-D Print-ing. In: Izdebska, J. & Thomas, S. (eds.) Print-ing on Polymers: Fundamentals and Applica-tions. Amsterdam, Elsevier, pp. 293-306.

26

Journal of Graphic Engineering and Design, Volume 7 (1), 2016. 27

Introduction

Tactile way of information perception by blind people dictates the appropriate requirements for quality of relief-dot labeling that must remains at consistently high level for a long period of use (European Union, 2004), which directly affects on the accuracy and correctness of the perception of the printed text (International Organization for Standardization, 2013). To ensure the durability of Braille symbols it is necessary to carry out deep analysis of process and conditions of relief prints’ operation, reveal the maximum complete set of influenc-ing factors on their quality in the use of blind readers. Many of these factors and relationships between them allow for their systematization and displaying by means of a directed graph. Designed model can be used for making decision and finding the best options for increas-ing the lifetime of printed products for blind people.

In the era of global computerization and automation the necessary decision in problem of quality control of relief-dot labeling is to develop a digital system of con-formity assessment Braille symbols with requirements of international standards and predict changes in the dot’s parameters during its tactile use by blind people.

Recent researches and publications

Research issues of ensuring blind people of printed tactile information devoted to works of local and for-eign scientists, such as: Beskov, Vakulich, Havenko, Kibirkshtis, Labetska (Havenko, Labetska & Khadzhy-nova, 2013; Havenko et al., 2013a; Havenko, et al., 2013b; Havenko et al., 2015a; Havenko et al., 2015b; Havenko, Labetska & Havenko, 2016), Mayik (Mayik, 2013), Onishchenko, Stepien, and others, where con-sidered problems of tyflo technology and equipment for the reproduction of information in Braille. How-ever, analysis of the literature and patent information indicates the lack of deep scientific developments regarding to research of the tactile information dura-bility, automation of control process of its quality.

The aim of research

The aim of research was to carry out simulation of evaluation process of Braille image’s durability by developing a digital system of control the conformity of Braille symbols parameters according to applicable international standards with defined its operation-al stability; to consider influencing factors on the quality of tactile labeling in their use of the blind.

Svitlana Havenko, Mykola Havenko, Marta Labetska Ukrainian Academy of Printing,Faculty of Publishing, Printing and Information Technologies, Lviv, Ukraine Corresponding author: Marta Labetska e- mail: marta_motyka@mail.ru

First recieved: 01.06.2016.Accepted: 27.06.2016.

abstract

The paper presents a model of factors that influence on the quality of relief-dot Braille inscriptions in their intensive use by blind people. It was constructed matrix of depending and distancing by using a directed graph. It allowed establish a hierarchy between studied criteria’s of impact the on durability of relief prints. It was developed a basic structure of digital system of assessment the conformity of Braille symbols to applicable international standards and definition of operational stability.

KEy WoRDSmodel, Braille, relief-dot image, digital system, quality.

simulation of durability the evaluation of relief-dot braille images

Preliminary report http://doi.org/10.24867/JGED-2016-1-027

28

Methods

Printing products, specially adapted for use by blind people (with relief-dot printed markings), character-ized by much shorter service life. Detailed analysis of influencing factors on the durability of Braille symbols allowed to isolate them in a set F={f1, f2, …, fn}, and select the most important ones (subset F1={f1, f2, …, f8}):

f1 – environmental impact; f2 – frequency of use; f3 – reader age; f4 – reader qualification; f5 – method of formation of relief elements; f6 – the type of carrier material of relief image; f7 – type of printing composition; f8 – additional coating of base material.

To represent the links between above listed factors used graph theory (Figure 1).

» Figure 1: The graph of the links between factors that influence on the durability of relief-dot products

Based on the building graph and using the known meth-od (Havenko & Gunko, 1996) it was constructed a matrix of dependency and distancing (Table 1,2) (Hlinenko & Suhonosov, 2003) for further installation of all levels of the hierarchy of factors (Stotsko, 2013) that influence on the durability of relief-dot markings and build the appro-priate model (Figure 2).

Table 1Matrix of dependency

f1 f2 f3 f4 f5 f6 f7 f8

f1 0 0 0 0 0 1 1 1

f2 0 0 0 0 1 1 1 1

f3 0 1 0 1 0 0 0 0

f4 0 1 0 0 1 0 0 0

f5 0 0 0 0 0 0 0 0

f6 0 0 0 0 1 0 0 0

f7 0 0 0 0 1 0 0 0

f8 0 0 0 0 1 0 0 0

Table 2Matrix of distancing

f1 f2 f3 f4 f5 f6 f7 f8

f1 1 0 0 0 1 1 1 1

f2 0 1 0 0 1 1 1 1

f3 0 1 1 1 1 1 1 1

f4 0 1 0 1 1 1 1 1

f5 0 0 0 0 1 0 0 0

f6 0 0 0 0 1 1 0 0

f7 0 0 0 0 1 0 1 0

f8 0 0 0 0 1 0 0 1

Results and discussion

Analysis of the received model allowed the claim that between all influencing factors on the durability of inscriptions in Braille the highest priority has their methods of formation and used materials. The next level is the frequency of use of relief-dot images by blind people. Further, on the less important levels of graphical model placed criteria’s of environmental impact and reader’s age.

Simulation of assessment process of the impact of a combination of factors on the quality of relief-dot char-acters during the use by blind people carried out by for-malization of relations between the criteria by means of a directed graph and corresponding matrix of distancing.

The received model can be used to obtain the numerical weight values of the studied factors and predict the dura-tion of the life cycle of printed products for blind people with high quality that meets international standards.

Simulation of process assessment the conformity parameters of Braille symbols to this requirements and durability of relief-dot images realized by developing a digital system, which algorithm is shown in Figure 3.

The principle of the digital system of automatic qual-ity control of relief markings during their operation is based on the simultaneous recording of Braille print and assessment of matches the geometric parameters of point with programmed requirements of existing international standards after each wear cycle. The result of this system is to establish the operational stability of relief prints by determining the level of damage.

Conclusions

It was presented the model of factors that influence on the Braille durability with the definition a hierarchy between them, which makes it possible to perform forecasting and improve operational stability of relief-dot marking.

Journal of Graphic Engineering and Design, Volume 7 (1), 2016. 29

» Figure 2: Graphic model of hierarchy between impact factors on the durability of Braille relief-dot elements

» Figure 3: Scheme of algorithm of developed digital system of Braille image’s quality evaluation during their operation

30

References

European Union. (2004) Directive 2004/27/EC of the European Parliament and of the Council. Official Journal of the European Union. L 136, 47, 34-57 Available from: http://eur-lex.europa.eu/legal-con-tent/EN/TXT/PDF/?uri=CELEX:32004L0027&-from=EN [Accessed 10th February 2016]

Havenko, S. & Gunko, S. (1996) Principles of mod-elling technical systems in printing. [Принципи моделювання технічних систем у друку]. Lviv, Ukrainian Academy of Printing.

Havenko, S., Labetska, M.& Khadzhynova, S. (2013) Printing methods to perform marking in Braille. [Poligraficzne metody wykonywa-nia nadruku alfabetem Braille’a]. Swiat druku.(07-08), 74-76. Available from: http://www.swi-atdruku.eu/Archiwum/Rok-2013/07-08-2013/Poligraficzne-metody-wykonywania-nedruku-al-fabetem-Braill-a [Accessed 16th February 2016]

Havenko, S., Kochubei, V., Labetska, M., Khadzhy-nova, S., Kibirkštis, E. & Venytė, I. (2015a) Ther-mal analysis of Braille formed by using screen printing and inks with thermo powder. Medziag-otyra. 2 (1), 52-56. Available from: doi: 10.5755/j01.ms.21.1.5702 [Accessed 3rd March 2016].

Havenko, S., Labetska, M. & Havenko, M. (2016) Mutual integration of computer information technology of Braille. [Presentation] Belarusian State Technological University, 12th February.

Havenko, S., Labetska, M., Kotmalova, O., Khadzhy-nova, S.& Havenko, M. (2013a) Using aroma printing and Braille to the packaging iden-tify. News of printing. 1 (370), 14-17.

Havenko, S., Labetska, M., Petryk, P. & Havenko, M. (2015b) The intensification of process of information perception by blind people with aroma labelled in Braille. Naukovedenie. 7 (2), 1-11. Available from: doi: 10.15862/46TVN215 [Accessed 10th March 2016].

Havenko, S., Labetska, M., Stepien, K., Kibirkštis, E. & Venytė, I. (2013b) Research of influencing fac-tors on the change of geometric parameters of Braille elements on self-adhesive labels. Mechan-ika. 19 (6), 716-721. Available from: doi: 10.5755/j01.mech.19.6.6016 [Accessed 6th March 2016].

Hlinenko, L. & Suhonosov, O. (2003) Funda-mentals of technical systems modelling. [Основы моделирования технических систем]. Lviv, Beskid Bit.

International Organization for Standardization. (2013) ISO 17351:2013. Packaging -Braille on pack-aging for medicinal products. Geneva, Inter-national Organization for Standardization.

Mayik, V. (2013) Analysis of the impact of the size and shape of Braille element on its integrated assessment. Scientific Notes. 2, 94-100.

Stotsko, Z. (2013) Simulation of technology systems. [Моделювання технологічних систем]. 2nd edition. Lviv, Publishing House Lviv Polytechnic.

12016June

7

Eye-tracking analysisof face observing and face recognition

Andrej Iskra, Helena Gabrijelčič Tomc

The analysis of ink jetprinted eco-font efficiency

Rastko Milošević, Uroš Nedeljković,Bojan Banjanin, Dragoljub Novaković,

Nemanja Kašiković

Development of tactile floor planfor the blind and the visually impaired

by 3D printing techniqueRaša Urbas, Matej Pivar, Urška Stankovič Elesini

Simulation of durability the evaluation ofrelief-dot Braille images

Svitlana Havenko, Mykola Havenko,Marta Labetska