Title :
Neural network structures for expression recognition
Author :
Ding, J. ; Shimamura, M. ; Kobayashi, H. ; Nakamura, T.
Author_Institution :
Dept. of Comput. & Math. Sci., Tohoku Univ., Sendai, Japan
Abstract :
The authors build up expression features models and then apply them to the network structures for expression recognition, focusing on how to determine the number of hidden nodes and initialize the weights. Moreover, the simulation results of the authors´ methods are provided to show the ability of the backpropagation neural networks for recognizing facial expressions.
Keywords :
backpropagation; image recognition; neural nets; backpropagation neural networks; expression features models; expression recognition; hidden nodes; neural network structures; Cost function; Eyes; Gold; Hospitals; Image processing; Mouth; Neural networks; Psychology; Shape;
Conference_Titel :
Neural Networks, 1993. IJCNN '93-Nagoya. Proceedings of 1993 International Joint Conference on
Print_ISBN :
0-7803-1421-2
DOI :
10.1109/IJCNN.1993.716813