Automated Detection of Human Emotion Jennifer Lee Quarter

Automated Detection of Human Emotion Jennifer Lee Quarter 1

Goal To be able to identify emotions using a low quality camera (webcam). Applications Human-Computer Interaction Alternate Reality Product Testing

Past Research Very good results About 80 -90% accuracy Generally have access to high quality cameras Two visual-based processes Marker based Anger Sadness Happiness Neutral Anger 0. 84 0. 08 0. 00 0. 08 Sadness 0. 00 0. 90 0. 00 0. 10 Happiness 0. 00 0. 98 0. 02 Neutral 0. 00 0. 02 0. 14 0. 84 Analysis of Emotion Recognition using Facial Expressions, Speech and Multimodal Information (2004)

Development Python (Open. GL, PIL) Read each image Head Shift Adjustments GUI Lighting Adjustments Feature Isolation Webcam Identify Markers Produce Tracking Image Analyze Tracking Image

Progress Python webcam integration Tracking Basic marker identification Basic marker tracking Head movement compensation Detailed marker tracking Identification Basic Identification Learning

Current Progress Facial tilt and movement test image.

Current Progress

Problems Success rate lower than desired Learning will improve this rate.