Development of algorithms for face and character recognition based on wavelet transforms, PCA and neural networks

In this paper we present a novel algorithms for face and character recognition using combination of wavelet transforms and principal component analysis (PCA). At first, face features are extracted using combination of Haar and Daubechies wavelet transform. Then obtained features are used for face recognition by PCA (eigenfaces). In the case of character recognition we use combination of wavelet transform and principal component analysis for character feature extraction. Then obtained extracted features are classified using multi-layer feed-forward neural networks. For each training character we use one neural network, which determines the confidence whether an input character is its prototype or not. The proposed algorithms give an effective performance of face and character recognition on noisy images and compete with state-of-the-art algorithms.