A Quantized/truncated Transform Domain technique (QTD) for fast facial recognition

A Quantized/truncated Transform Domain technique (QTD) is presented in this paper for facial recognition, suitable particularly for large data bases. The algorithm has attractive properties with respect to storage requirements and computational complexity in both the training and testing modes. The new algorithm is applied to both ORL and Yale data bases using Discrete Cosine Transform (DCT). The experimental results confirm the significant reduction in the storage and computational requirements while maintaining a high recognition accuracy rate.