Dimensionality Reduction using SOM based Technique for Face Recognition

Unsupervised or Self-Organized learning algorithms have become very popular for discovery of significant patterns or features in the input data. The three prominent algorithms namely Principal Component Analysis (PCA), Self Organizing Maps (SOM), and Independent Component Analysis (ICA) have widely and successfully been used for face recognition. In this paper a SOM based technique for dimensionality reduction has been proposed. This technique has also been successfully used for face recognition. A comparative study of PCA, SOM and ICA along with the proposed technique for face recognition has also been given. Simulation results indicate that SOM is better than the other techniques for the given face database and the classifier used. The results also show that the performance of the system decreases as the number of classes increase.