Human Face Recognition
|Training and Testing of Neural Networks|
Performance Analysis and Discussions
Training and Testing of Neural Networks
Two neural networks, one for PCA based classification and the other for LDA based classification are prepared. ORL face database is used for training and testing. The training is performed by n poses from each subject and the performance testing is performed by 10-n poses of the same subjects. After calculating the eigenfaces using PCA the projection vectors are calculated for the training set and then used to train the neural network. This architecture is called PCA-NN. Similarly, after calculation of the fisherfaces using the LDA, projection vectors are calculated for the training set. Therefore, the second neural network is trained by these vectors. This architecture is called LDA-NN. Figure 9 shows the schematic diagram for the neural network training phase. When a new image from the test set is considered for recognition, the image is mapped to the eigenspace or fisherspace. Hence, the image is assigned to a feature vector. Each feature vector is fed to its respective neural network and the network outputs are compared.
Figure 9. Training phase of both Neural Networks
In this chapter, two face recognition systems, the first system based on the PCA preprocessing followed by a FFNN based classifier (PCA-NN) and the second one based on the LDA preprocessing followed by another FFNN (LDA-NN) based classifier, areintroduced. The feature projection vectors obtained through the PCA and LDA methods are used as the input vectors for the training and testing of both FFNN architectures. The proposed systems show improvement on the recognition rates over the conventional LDA and PCA face recognition systems that use Euclidean Distance based classifier. Additionally ,the recognition performance of LDA-NN is higher than the PCA-NN among the proposed systems.
Figure 10. Recognition rate vs. number of training faces
Want To Know more with
Contact for more learning: webmaster@freehost7com
The contents of this webpage are copyrighted © 2008 www.freehost7.com
All Rights Reserved.