Name
Abstract | ||
|---|---|---|
A fully automated, multi-stage architecture for emotion recognition is presented. Faces are located using a tracker based upon the ratio template algorithm [1]. Optical flow of the face is subsequently determined using a multi-channel gradient model [2]. The speed and direction information produced is then averaged over different parts of the face and ratios taken to determine how facial parts are moving relative to one another. This information is entered into multi-layer perceptrons trained using back propagation. The system then allocates any facial expression to one of four categories, happiness, sadness, surprise, or disgust. The three key stages of the architecture are all inspired by biological systems. This emotion recognition system runs in real-time and has a range of applications in the field of human-computer interaction. |
Year | DOI | Venue |
|---|---|---|
2003 | 10.1007/3-540-44887-X_15 | AVBPA |
Keywords | Field | DocType |
different part,facial part,emotion recognition,multi-stage architecture,facial expression,direction information,key stage,biological system,emotion recognition system,real-time emotion recognition,human-computer interaction,social interaction,optical flow,real time,human computer interaction,human interaction,biological systems,non verbal communication,social intelligence,multi layer perceptron,back propagation | Facial recognition system,Sadness,Pattern recognition,Computer science,Facial expression,Artificial intelligence,Biometrics,Backpropagation,User interface,Perceptron,Optical flow,Distributed computing | Conference |
Volume | ISSN | ISBN |
2688 | 0302-9743 | 3-540-40302-7 |
Citations | PageRank | References |
4 | 1.00 | 10 |
Authors | ||
2 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Keith Anderson | 1 | 190 | 11.01 |
| Peter W. McOwan | 2 | 1282 | 55.89 |