Sorted evolutionary strategy based SOFM used for vector quantization

We present a sorted evolutionary strategy based self-organizing feature map (SOFM) algorithm to improve the efficiency of vector quantization. The image samples are sorted according to the human vision sensitivity to ensure an optimal vision effect under the precondition of the globe minimum error. A similarity evaluation about code vector is introduced to the evolutionary algorithm to guarantee the variety of the code vector and the adaptability to the image. Experimental results show that the higher adaptability of codebook and better quality of reconstructed image.