Emotion Recognition Based on Infrared Facial Image

1

• Emotion Recognition Based on Infrared Facial Image

Peng Yang Rutgers University Supervisor Jinjun
Wang
2
Motivation Safe drive!!!
3
Motivation

• Goal of this project
• Provide alarm to driver based on drivers emotion
  status
• Challenges
• Temperature changes under the same emotion.
• Different subjects have different temperature
  level.
• The temperature change trends are different.

4
Background

• Preliminaries
• Related Work
• Expression recognition using IR image (Masood
  2008).
• Face recognition in the thermal infrared Spectrum
  (Pradeep 2007).

5
Project Overview
Happy
Angry
Fatigue
Neutral
Sadness
6
Project Overview
Infrared camera
Image
Face detection
Face crop
Feature extraction
Classifier
Result
7
Project Description
Features
Classifier
PCA
Boosting
Gabor
Regression
Mean of Patch
SVM
VLBP
8
Experiment

• Experimental results--if no experiment, use this
  slide for more project description (avoid table
  if possible)

9
Conclusion, Future Work, Feedbacks

• Emotion status shifts along the time.
• Different between subjects are very large.
• The data set are too noisy.
• Physiology scientists help will be very
  helpful.