Abstract: In this study, a face recognition system using PCA on eigenimages is proposed. This system automatically detects the eyes region from a face image using the wavelet transform. These detected eyes are separated from the face image then eigeneye decomposition is performed. A new classification based on the least squares estimate to improve recognition accuracy is also proposed. Good improvement in classification is achieved which outperforms traditional PCA on the whole face especially when taking into accoun the substantial reducation in the computational complexity of the proposed algorithm.