A Feature-Based Face Recognition Approach Using Gabor Wavelet Filters Cascaded with Concurrent Neural Modules

We propose a face recognition model consisting of the following stages: facial feature localization (23 essential points, corresponding to eyes, mouth, nose, and face boundary); feature representation by Gabor wavelet based filtering (GWF); dimensionality reduction using principal component analysis (PCA); neural classification using concurrent self-organizing maps (CSOM). For the ORL face database, we obtain a recognition score of 96%.