Color image compression using visual pattern vector quantization combined to an adaptive block scanning

Abstract

This paper demonstrates a new method for color image compression. The compression used here is a lossy color image compression. In this paper, a color image is compressed and reconstructed using vector quantization. After the preprocessing step, vector quantization encoder is applied followed by block based DCT transform, quantization, encoder and the reverse part such as decoding, inverse DCT, dequantization, and vector quantization decoding followed by YCbCr transformation are performed. These are performed to obtain better quality and compression ratio. In order to obtain better compression ratio, four types of adaptive scanning such as horizontal, vertical, zigzag and Hilbert are performed. The proposed method achieves better performance than JPEG in terms of visual quality and compression ratio.