Headshot My research is at the intersection of psychology and computer science. I am interested in designing hardware and algorithms for sensing human behavior at scale, and in building technologies that make life better. Applications of behavior sensing I am most excited about are in: understanding mental health, improving online learning and designing new connected devices (IoT).
contact: djmcduff (at) media.mit.edu ·  twitter: @danmcduff
google scholar: profile
linkedin: profile
links:
Publications
CV
Press
Videos
Research Stat. Teaching Stat.


Research Interests: HCI, Emotion, Computer Vision, Bio-Sensors, Crowdsourcing, Machine Learning, Psychology

Daniel is Director of Research at MIT Media Lab Spin-Out Affectiva and leading collection and analysis of the world's largest database of human emotion data. Daniel McDuff received his Ph.D. from the MIT Media Lab while working in the Affective Computing group. McDuff received his bachelor's degree, with first-class honors, and master's degree in engineering from the University of Cambridge. Prior to joining the Media Lab, he worked for the Defense Science and Technology Laboratory (DSTL) in the United Kingdom. He is interested in using computer vision, wearable sensors and machine learning to enable the automated recognition of affect. He is also interested in technologies and interfaces for remote measurement of physiology for which his work has received an award from Popular Science magazine as one of the top inventions in 2011.

Currently much of my work is focused on using deep learning to analyze the largest set of facial expression videos in the world which we have collected at Affectiva.

The technology Daniel has helped develop is being commercialized by a number of companies including start-ups: Affectiva and Cardiio.

His work has been reported in many publications:

Smiley Smiley Smiley Smiley Smiley
Smiley Smiley Smiley Smiley Smiley

A number of videos of my work are available here.


Daniel McDuff Director of Research
Affectiva
Cambridge
MA 02139
USA


SXSWSXSW Awards 2014. Winner with Pavlov Poke!

Updates:
April 2016: I will be speaking at the Deep Learning in Healthcare Summit in London (April 7-8th).
Feb 2016: Our CHI Note "COGCAM: Contact-free Measurement of Cognitive Stress During Computer Tasks with a Digital Camera" received a CHI honorable mention.
October 2015: I spoke at the WIRED UK Conference in London (October 16th). Video available here.
September 2015: I spoke at TEDXBerlin TEDxBerlin in Germany (September 6th). Video available here.
July 2015: I spoke at the Sentiment Analysis Symposium in New York (July 15-16th).
June 2015: I was selected as a 2015 Wired Innovation Fellow


Physiology ·  Crowdsourcing & Emotions ·  Interfaces ·  Facial Expressions ·  Other Projects · 

Ph.D. Thesis:


My Ph.D. thesis was the first work to present large-scale collection and analysis of emotional responses in uncontrained real-world settings (using people's webcams to capture their facial expressions and physiology).

Smiley Affective Storytelling (Ph.D. Thesis)
Daniel McDuff
Automatic Measurement of Story Effectiveness from Emotional Responses Collected over the Internet
thesis



Remote Measurement of Physiology:


Methods for measuring physiological signals without contact with the human body. Most of this work focuses on using ordinary RGB webcams to capture heart, breathing and HRV measurements. I also curate a webpage with resources related to iPPG measurement.

CardioCam CardioCam
Ming-Zher Poh, Daniel McDuff and Rosalind Picard
pdf one ·  pdf two ·  video
3RD PRIZE CIMIT AWARD FOR PRIMARY HEALTHCARE - $50K PRIZE
CardioCam Multi-spectral Remote Physiology Measurement
Daniel McDuff, Sarah Gontarek and Rosalind Picard
Improvements Using a Five-Band Digital Camera
pdf
Medical Mirror The Medical Mirror
Ming-Zher Poh, Daniel McDuff, Rosalind Picard
A Medical Mirror for Non-contact Health Monitoring
Siggraph 2011, Vancouver.
pdf
CardioCam Dectecting Cognitive Load with a Camera
Daniel McDuff, Sarah Gontarek and Rosalind Picard
EMBC2015 Paper and Presentation
pdf
CHI 2016, San Jose, Note
pdf
CHI Honorable Mention
Review Review of Remote Physiological Measurement Methods
Daniel McDuff
In collaboration with the Air Force Research Lab
EMBC2015
pdf ·  webpage and resources
Multi-Camera Measurement of Physiology
Daniel McDuff
Dynamically Fusing Information from Multiple Imagers
More coming soon.



Wearable Physiological Measurent:


Measuring physiology via subtle motions of the body using just a smartphone or watch.

BioWatch BioWatch
Javier Hernandez, Daniel McDuff and Rosalind Picard
pdf
Nominated for best paper award - Pervasive Health
BioInsights BioInsights
Javier, Hernandez, Daniel McDuff and Rosalind Picard
Extracting Personal Data from "Still" Wearable Motion Sensors
pdf
BEST POSTER AWARD - Body Sensor Networks
BioWatch BioPhone
Javier Hernandez, Daniel McDuff and Rosalind Picard
pdf
Nominated for best student paper award - IEEE EMBC



Analyzing Large-scale Observational Emotion Measurements:


I have been leading collection of the WORLD'S LARGEST DATASET of emotion responses.
Currently it contains over 4 Million videos from nearly 80 countries.
This data allows us to understand and model human behaviors in a new way.

BioWatch Cross-Cultural Expressions
Daniel McDuff, Jeffrey Girard and Rana el Kaliouby
Observational evidence of cultural differences across 100,000s of people.
More coming soon.
BioWatch Gender Differences in Expressions
Daniel McDuff, Evan Kodra, Marianne LaFrance and Rana el Kaliouby
Smiling differences between thousands of men and women measured in everyday settings.
More coming soon.



thesis_paperPredicting Purchase Intent from Facial Responses to Ads
Daniel McDuff, Rana El Kaliouby, Jeffrey F. Cohn, Rosalind Picard
Transaction on Affective Computing
pdf
Affect Mirror AMFED Facial Expression Dataset
Daniel McDuff, Rana El kaliouby, Thibuad Senechal, May Amr, Jeffrey F. Cohn and Rosalind Picard
CVPRW 2013, Portland, OR.
dataset ·  pdf
Smiles-in-the-wild Predicting Online Video Liking from Facial Responses
Daniel McDuff, Rana el Kaliouby, David Demirdjian and Rosalind Picard
Predicting Online Media Effectiveness Based on Smile Responses Gathered Over the Internet.
IEEE Conference on Automatic Face and Gesture Recognition 2013, Shanghai.
pdf
WINNER OF BEST STUDENT PAPER AWARD
Automatic Measurement of Ad Preferences from Facial Responses Gathered Over the Internet.
Image and Vision Computing, 2014.
pdf
Dial Automatic Measurement of Media Preferences (work completed at Affectiva)
Evan Kodra, Thibaud Senechal, Daniel McDuff and Rana el Kaliouby
Automatic Measurement of Media Preferences from Facial Behavior
IEEE Conference on Automatic Face and Gesture Recognition 2013, Shanghai.
pdf
Smiles-in-the-wild Analyzing Facial Expressions Over the Web (with Forbes.com)
Daniel McDuff, Rana el Kaliouby and Rosalind Picard
Crowdsourcing Facial Responses to Online Videos.
IEEE Transaction on Affective Computing 2012
pdf ·  site



Affective Interfaces:


Understanding affective data can be difficult.
I am designing new ways to interact with this data to help people reflect on their emotions.

Affective_Billboard Affective Billboard
Daniel McDuff, Yadid Ayzenberg and Rosalind Picard
Video - coming soon
Affect Mirror Inside-Out: Reflecting on your Inner State
Javier Hernandez Rivera, Daniel McDuff, Rich Fletcher and Rosalind Picard
Pervasive Computing 2013, San Diego, CA.
pdf
3RD PRIZE OLYMPUS/MIT HACKATHON 2012
AffectAura AffectAura: An Intelligent System for Emotional Memory (work completed at MSR)
Daniel McDuff, Amy Karlson, Ashish Kapoor, Asta Roseway and Mary Czerwinski
An emotional prosthetic that allows users to reflect on their emotional states over time.
CHI 2012, Autstin, TX.
pdf
AutoEmotive AutoEmotive
Javier Hernandez, Daniel McDuff, Judith Amores and Xavier Benavides
Webpage
WINNER OF MIT/VW/IDEO DATA DRIVEN HACKATION 2014 - $5K PRIZE



Healthcare Applications:
PAD VISION
Vascular disease screening.
Daniel McDuff, Ann DeBord Smith, Becca Scully, Louis Nguyen
More coming soon.
Suicide Risk Assessement
Steven Vannoy, Daniel McDuff, Rana el Kaliouby
More coming soon.



Facial Expression Analytics:


Detecting and interpreting facial expressions using computers.
These projects concern the development of automated facial coding software.

FG2015_b Group Sparsity for AU Detection
Daniel McDuff, Yale Song, Deepak Vasisht, Ashish Kapoor
IEEE AFGR 2015
pdf
FG2015_a Ultra Large-Scale Expression Clustering
TJ Vandal, Daniel McDuff, Rana El kaliouby
IEEE AFGR 2015
pdf
ESOMAR Do Emotions in Ads Drives Sales?
Daniel McDuff, Rana El kaliouby, Evan Kodra and Laurent Larguinet
ESOMAR Congress 2013
pdf
BEST METHODOLOGY PAPER AWARD NOMINEE
BEST CASE HISTORY PAPER AWARD NOMINEE
BEST OVERALL PAPER AWARD NOMINEE
Election Measuring Emotions During the 2012 Presidential Debates
Daniel McDuff, Rana El kaliouby, Evan Kodra and Rosalind Picard
ACII 2013, Geneva, Switzerland.
pdf
ESOMAR Facial AU Detection using Active Learning and an Efficient Kernel Approximation
Thibaud Senechal, Daniel McDuff and Rana El kaliouby
IEEE ICCV Workshops 2015
pdf



Other Projects:
AutoEmotive time_and_space
A data visualization for the CTL Contest
Daniel McDuff
AWARD WINNER
Webpage
Thumbs Down Pavlov Poke
A Shocking Solution to Facebook Addiction
Rob Morris and Daniel McDuff
SXSW INTERACTIVE AWARDS WINNER 2014
website
Crowdsource Crowdsourcing and Emotions
Review
Rob Morris and Daniel McDuff
pdf
FAUS Is that smile real or fake?
M. Ehsan Hoque, Daniel McDuff and Rosalind Picard
Exploring Temporal Patterns in Classifying Frustrated and Delighted Smiles.
IEEE Transaction on Affective Computing, 2012
pdf
Narrative Emotion and Narratives
Burzo, M., McDuff, D., Mihalcea, R., Morency, L.P., Narvaez, A. and Perez-Rosas, V.
Towards Sensing the Influence of Visual Narratives on Human Affect.
pdf
Know Your Exit Know Your Exit - Crowdsourced Music Project
Robert Morris, Daniel McDuff and Rosalind Picard
SXSW INTERACTIVE AWARDS FINALIST 2012
site
Acume Acume - Data Viz Toolkit
Daniel McDuff, Rana el Kaliouby, Karim Kassam and Rosalind Picard
A novel visualization tool for understanding facial expression and gesture data.
pdf ·  software ·  video
FAUS Facial Action-Unit Spectrograms and Predicting Valence
Daniel McDuff, Rana el Kaliouby, Karim Kassam and Rosalind Picard
Affect Valence Inference From Facial Action Unit Spectrograms.
pdf
healthcare Migraine Trakz
Third Prize in the MGH-Samsung Healthcare Hackathon.
news article