J U N E 2 0 1 8 7C O M P U T E R 0 0 1 8 - 9 1 6 2 / 1 8 / $ 3 3 . 0 0 © 2 0 1 8 I E E E P U B L I S H E D B Y T H E I E E E C O M P U T E R S O C I E T Y J U N E 2 0 1 8 7

EDITOR RON VETTER University of North Carolina Wilmington;

vetterr@uncw.eduSPOTLIGHT ON TRANSACTIONS

W hat would fictional character Sheldon Cooper from CBS’s television series “The Big Bang Theory” do if he wanted to study the differences in autistic individuals’

facial expression of emotions? He might be intrigued by rapid improvements in the field of affective computing, in which computer-supported behavior analysis—by means

of automated facial, bodily, or vocal expression analysis—promises to provide deeper insight into differ-ences across various neurodevelop-mental and other disorders.

For example, Tanaya Guha, Zhao-jun Yang, Ruth B. Grossman, and Shrikanth S. Narayanan’s recent study “A Computational Study of Expressive Facial Dynamics in Chil-dren with Autism” (IEEE Trans. Affec-tive Computing, vol. 9, no. 1, 2018, pp. 14–20) uses affective computing tech-niques to follow up in more detail on the difference between the facial ex-pressions of young typically develop-ing (TD) individuals and their coun-terparts on the autism spectrum.

The authors used motion cap-ture to track the 32 “most critical” facial points by reflective markers (including four stability markers)

affixed across the face of the participants during data collection. Twenty children with high-functioning au-tism (HFA) and 19 TD peers (all between 9 and 14 years of age and mostly male, in keeping with autism’s common prevalence distribution) mimicked expressions accord-ing to video stimuli selected from the broadly used Mind Reading corpus in the Ekman “Big Six” emotion classes.

What Affective Computing Reveals about Autistic Children’s Facial Expressions of Joy or FearBjörn Schuller, University of Augsburg and Imperial College London

This installment of Computer’s series

highlighting the work published in IEEE

Computer Society journals comes from IEEE

Transactions on Affective Computing

8 C O M P U T E R W W W . C O M P U T E R . O R G / C O M P U T E R

SPOTLIGHT ON TRANSACTIONS

The authors divided the face region into three macro-areas of equal height from forehead to chin—upper, mid-dle, and lower face. Mean square error across tracked facial points in a region was then used to investigate complex-ity and (dis-)similarity of expressions across the two groups.

Guha and her colleagues found sig-nificant differences between the HFA and TD participants for disgust and sadness in all three facial areas, for joy in the upper and lower face, and for surprise in the lower face—but none for anger or fear. The observed dissimilarities were marked by lower complexity of the facial expression display among the children with HFA. The study further found that the chil-dren with HFA showed the highest

complexity, and thus most natural ex-pression, in the cheek region, in con-trast to reduced complexity of move-ments in the eye region. They reason that the latter might be explained by the observation that individuals with autism tend to avoid looking at the eye region of others’ faces. This dif-ference in complexity across regions might explain the common description of autistic facial emotional display as atypical and awkward.

Computer analysis has yet to unleash its full potential in supporting behavioral and

more general psychological studies. The authors stress the limitations of their study, given its smaller sample

size and mimicked emotions. Yet, in light of the recent melding of artificial (emotional) intelligence with increas-ingly “big(ger) data” resources that can be handled even under adverse in-the-wild conditions, it will be excit-ing to see what new insights affective computing will bring as a supporting discipline feeding into psychology, medicine, and much more.

BJÖRN SCHULLER is a professor of

embedded intelligence for healthcare

and wellbeing at the University of

Augsburg, and a reader in machine

learning at Imperial College London.

Contact him at schuller@ieee.org.

IEEE Software offers pioneering ideas, expert analyses, and thoughtful insights for software professionals who need to keep up with rapid technology change. It’s the authority on translating software theory into practice.

www.computer.org/ software/subscribe

ARCHITECTURAL DESIGN PRINCIPLES // 15

WHY LARGE IT PROJECTS FAIL // 117

SOFTWAREQUALITY

WWW.COMPUTER.ORG/SOFTWARE

JULY/AUGUST 2016

IEE

E S

OF

TW

AR

E

July/August 2016

SO

FT

WA

RE

QU

AL

ITY

V

olu

me 33 N

um

ber 4

TINY PROGRAMMING TOOLS // 24

REQUIREMENTS & SOCIAL RESPONSIBILITY // 109

WWW.COMPUTER.ORG/SOFTWARE

JANUARY/FEBRUARY 2016IE

EE

SO

FT

WA

RE

January/February 2016

TH

E F

UT

UR

E O

F S

OF

TW

AR

E E

NG

INE

ER

ING

V

olu

me 33 N

um

ber 1

CODE CLARITY // 22

SOFTWARE ON A COMET // 81

WWW.COMPUTER.ORG/SOFTWARE

MARCH/APRIL 2016

IEE

E S

OF

TW

AR

E

March/A

pril 2016 B

IG D

ATA

V

olu

me 33 N

um

ber 2