An online image compression algorithm using singular value decomposition and adaptive vector quantization

In this paper we propose novel strategies for reducing the practical limitations of the Singular Value Decomposition (SVD) for image compression. First, the computational cost is reduced by computing only a limited number of eigenvectors. Second, an online Adaptive Vector Quantization (AVQ) method is used to achieve a low bit rate. Simulation results show that the proposed algorithm is better than the discrete cosine transform (DCT) and previous hybrid VQ-SVD in terms of distortion, bit rate, and image quality.