facial expression classification in still images
play

Facial expression classification In still images Angel Gutirrez, - PowerPoint PPT Presentation

Facial expression classification In still images Angel Gutirrez, Montse Pards Introduction Face detection Contour detection Expression estimation System implementation - DEA Results Conclusions Surprise Basic expressions (Ekman i


  1. Facial expression classification In still images Angel Gutiérrez, Montse Pardàs

  2. Introduction Face detection Contour detection Expression estimation System implementation - DEA Results Conclusions

  3. Surprise · Basic expressions (Ekman i Friesen – 1971) Sad Neutral INTRODUCTION Anger Joy

  4. INTRODUCTION Objective • To develop an automatic system which can recognize a facial expression in a still image. Applications • Monitor for drivers • Stress detection • Facial coder • Entertainment/ Games • etc.

  5. INTRODUCTION Generic scheme. Methods. INFORMATION FACE FACIAL EXPRESSION INPUT EXTRACTION EXPRESSION DETECTION DECISION IMAGE · Gabor Wavelets filters · Viola and Jones · Hidden Markov Models · Active appearance models · Neural Networks · Region based · Dense motion fields · Probabilistic models · Feature points tracking

  6. EXPRESSION FACE FEATURES DECISION DETECTION DETECTION

  7. FACE DETECTION Viola and Jones: · Fast search · Robust to background · Uses local texture features · Trains classifiers for face/non-face classes · Uses a cascade of classifiers structure (Adaboost) Region-based: · Refines the face detection to obtain a better initizialization

  8. FACIAL FEATURE CONTOUR DETECTION Model based method Active Shape Models Active Appearance Models

  9. FACIAL FEATURE CONTOUR DETECTION Based on Active Appearance Model software implemented by Stegmann* * M. B. Stegmann, B. K. Ersbøll, R. Larsen, FAME - A Flexible Appearance Modelling Environment, IEEE Transactions on Medical Imaging, vol. 22(10), pp. 1319-1331, Institute of Electrical and Electronics Engineers (IEEE), 2003 - Facial feature contours are represented by a 58 points model

  10. FACIAL EXPRESSION INPUT CLASSIFIER KNOWN DATA SYSTEM OUTPUT Class 1, Class 1, Class 2, Class 2, ... LABELED ... DATABASE Class N. Class N. Bayesian framework with probabilistic model: Mixture of multivariate gaussians, trained with EM for each class

  11. · 192 labeled images. RESULTS Database

  12. RESULTS

  13. RESULTS TEST 1 TEST 1 EXPRESSION FACE FEATURES DECISION DETECTION DETECTION 95,47% H A N Su Sa H 96% 0% 0% 4% 0% A 0% 98% 0% 0% 2% 95,47 % N 0% 0% 94% 6% 0% Su 4% 0% 3% 93% 0% Sa 0% 0% 3% 0% 97%

  14. RESULTS TEST 2 TEST 2 EXPRESSION FACE FEATURES DECISION DETECTION DETECTION 84,66% H A N Su Sa H 96% 0% 2% 2% 0% A 6% 79% 4% 4% 8% 84,66 % N 0% 3% 69% 13% 16% Su 0% 0% 4% 96% 0% Sa 0% 3% 14% 0% 83%

  15. CONCLUSIONS · Automatic system for facial expression detection in 3 stages: EXPRESSION FACIAL FEATURE FACE DETECTION CLASSIFICATION CONTOURS DET. ·Correct classification rate 85 %, with 5 classes . · Only frontal faces. Problems with facial hair and sometimes with glasses

Recommend


More recommend