SAMBA Face system Gloss Extraction - … · © Copyright Bossa Nova Technologies, LLC 2015 Date of study : May 2005 Product studied : Anti-shine skin lotion for face application

© Copyright Bossa Nova Technologies, LLC 2015

BOSSA NOVA TECHNOLOGIES

SAMBA Face system – Gloss Extraction


Located in Los Angeles (Culver City), CA, USA

LLC (Limited Liability Company) founded in 2002

Small business of 7 people specialized in optics, electronics, imaging and software development (2 PhDs, 2 MSc, 1 Engineers, 2 Technicians)

Manufacturer of testing equipment : laser ultrasonic inspection equipment, polarization cameras and systems for the cosmetic testing

Provides research for NASA, NSF, DoD and corporate clients

Well-known provider of imaging solutions of Non-Destructive Testing for various industries

Company Overview


• Laser Ultrasonic receiver and system (TEMPO, QUARTET, QUARTET LTG)

• Polarization cameras (SAMBA and SALSA)

• Cosmetic testing systems (SAMBA Face, SAMBA Hair, RUMBA and MAMBO)

Products line and R&D topics


Polarization of light depends on scattering event

Vertical Polarizer SPECULAR

Polarization camera

Skin or hair

Pigments

DIFFUSED

Polarization preserved

Random Polarization

Light source


Diffused and specular light separation with SAMBA camera

SAMBA CAMERA

Liquid Crystal

CC

D

Ele

ctr

on

ics

Diffused

Specular

Lens

Frame 1

Polarization imaging is in real-time & at video rate !

Frame 2

Crossed image

Parallel image


FRAME 1: Parallel Polarization

100% Specular + 50% Diffused ≈ Conventional Image

Example of polarization images



FRAME 2: Crossed Polarization

0% Specular + 50% Diffused ≈ Color image



Gloss/Shine Image =

Specular Image =

Frame 1 -Frame 2 =

Parallel - Diffused


SAMBA Face System hardware

SAMBA camera

Polarized white LED ring

Rotating chinrest

Hair system

Laptop PC Included – Pre-loaded SAMBA Face software


SAMBA Face Software architecture


Acquisition of images

Software keeps track of many studies, panelists and visits

Green target overlay + Ghost image feature to help reposition panelist between visits


Many possible camera views

Thanks to a rotating chin-rest


Regions Of Interest (ROIs)

User selects a ROI from a library to chose the area where the analysis will be done.

Multiple ROIs can be applied to a same image.

Data extraction in the ROI

Bare lips After lipstick application

Gloss increase

Gloss histogram: pixel count (y-axis) for each gloss intensity (x-axis)


Review and export images for reports

Crossed

Gloss

Gloss degree

Contrast Parallel

Gloss mapping


Applications Examples

Rinse-off products on forearms

Lipsticks

Skin translucency

Sebum prevention


Rinse-off products on forearms 1/7 Setup

Color SAMBA camera

acquiring two states of

polarization at video rate

Polarized illumination

Arm

+

Hardware [custom version]

Software

Visual Appearance Study Software

(camera calibration, export to excel,

ability to handle many studies,

panelists and visits...)


Rinse-off products on forearms 2/7 Image acquisition

Acquisition menu

The software helps the panelist reposition him/herself thanks to

a ‘ghost image’ that can be superimposed to the live image

Visit 02: bare skin

Visit 03: after application

and rinse-off of product A

on bare skin

Visit 04,05 & 06: after

application and rinse-off of

product B on bare skin


Rinse-off products on forearms 3/7 Gloss image calculation

For each visit: 1 Parallel and 1 Crossed image acquired

Parallel image = Specular light

(mirror-like) + 50% diffused light

Crossed image= 50% diffused light

- = Gloss image computed =

specular light only


Rinse-off products on forearms 4/7 Visual assessment

Bare skin Product A

Gloss increase with Product A Gloss increase with Product B


Rinse-off products on forearms 5/7 Choice of the ROI

The same ROI is

applied to all visits


Rinse-off products on forearms 6/7 Results: histograms on the ROI

Nu

mb

er

of o

ccu

rren

ce

in R

OI

Pixel value

Gloss increase

+160%

Gloss increase

Same ROI size

8bits/12bits scale available


Nu

mb

er

of

occ

urr

en

ce

in

RO

I

Pixel value

Crossed histograms for all visits Gloss histograms for visits 4,5 and 6 (product B)

Repeatability of ≈4%

Pixel value

Crossed signal (skin color)

remain the same all along the

visits

Rinse-off products on forearms 7/7 Repeatability


Acquisition menu

(take pictures and handle many studies, panelists, visits...)

Visit 16: bare lips

Visit 22: after application

of a lipstick

Lipsticks 1/4


Bare lips After application

Gloss increase after application

Lipsticks 2/4


Lipsticks 3/4

The same ROI is applied to all visits

Nu

mb

er

of o

ccu

rren

ce

in R

OI

Pixel value

+88%

Gloss increase

Same ROI size

Gloss increase


Nu

mb

er

of

occ

urr

en

ce

in

RO

I

Pixel value

Crossed histograms (before in blue

and after application in red)

Crossed signal (lips color) remains the

same. We observe only an effect due to

an increase of shine.

Crossed image before Crossed image after

Lipsticks 4/4


Skin translucency 1/3 Based on the publication of Akira Matsubara

Personal Beauty Care, Research & Development, Procter & Gamble Japan 1-17 Koyocho-Naka, Higashi-nada-ku, Kobe, 6580032, Japan

Objective: translate visual sensory perception into a measurable numerical parameter issued from polarization-based images. Definition of translucency: is a holistic term that involves the integration of skin color, glow and texture Method: -30 judges visually evaluated facial images of 32 female subjects to determine translucent skin and dull skin clinically

Table of visual assessment categories

from the publication of Akira Matsubara, P&G, 1-17 Koyocho-Naka, Higashi-nada-ku, Kobe, 6580032, Japan

-Specular rate (or Gloss Degree) defined from the specular and non-specular images was calculated as a numerical index for skin translucency. Relevancy of gloss degree to visual translucency was examined statistically.

Specular component

Total appearance Specular rate (gloss degree) =


Skin translucency 2/3


Results of the visual assessment

32 people classified in 3 categories: translucent - average - dull



Mean of the blue channel (skin tone) and its standard deviation (skin roughness)

is greater for dull skin

Results with SAMBA

Skin translucency 3/3


Date of study : May 2005 Product studied : Anti-shine skin lotion for face application Nb of panelists chosen for the study : 20 Purpose : Quantify the efficacy of product during the 12 hours time period following product application Principle of study : -After washing the skin, the product is applied on only one side of the panelist forehead, the other side is left untreated -Gloss measurements were taken almost every hour The evolutions of the gloss on both sides of the face are then compared to evaluate the influence of the product

Sebum prevention 1/6

Previous version of SAMBA hardware



1

10

100

1000

1 26 51 76 101 126 151 176 201 226 251

gloss after x hours ROI1

gloss after x hours ROI2

Treated side

Untreated side

Parallel Crossed

Gloss mapping


Panelist #1, gloss mapping images of the forehead

After cleaning At application T0 After 1 h

After 2 h After 3 h After 4 h After 8 h After 12 h

High gloss Low gloss

Gloss Scale

Untreated side



Sebum prevention 4/6 Histograms Castay Left

1

10

100

1000

0 10 20 30 40 50 60 70 80 90 100 110 120 130

Time Gloss increase(+) or

Decrease(-)

app 0%

1h -5%

2h -15%

3h -20%

4h -5%

8h -10%

12h +5%

Gloss results for treated side (panelist #1) Histograms Castay Right

1

10

100

1000

0 20 40 60 80 100 120 140

Time Gloss increase(+) or

Decrease(-)

app 0%

1h +25%

2h +14%

3h +17%

4h +33%

8h +26%

12h +40%

Gloss results for untreated side



Castay

-30

-20

-10

0

10

20

30

40

50

0 2 4 6 8 10 12

Time (hour)

Glo

ss v

ari

ati

on

(%

)

left

right

Panelist #1

treated

untreated

On this

panelist,

the gloss of the

treated side

does not

increase !

Product

application


Time Average gloss

difference

Application Reference

1h 4.1%

2h 6.8%

3h 6.3%

4h 6.0%

8h 13.3%

12h 13.3%

Average gloss

difference between

treated side and

untreated side

• The study shows that the

evolution of the gloss on the

untreated and treated sides

are quite different.

• On an average of 20 people,

the study reveals that the

gloss of the untreated side

increases much faster than the

gloss of the treated side.

• The study reveals that the

product slowed the increase of

gloss. After 3 hours, thanks to

the product , the face appears

6 % less glossy, and after 12

hours, it appears 13 % less

glossy.

Conclusion



Thank you!

Documents

SAMBA Face system Gloss Extraction - … · © Copyright Bossa Nova Technologies, LLC 2015 Date of study : May 2005 Product studied : Anti-shine skin lotion for face application