This paper concerns the compression of image data. A method is proposed to detect the edge components from images and to approximate the remaining planec omponents by polynomials. The coefficients of the polynomial is vector quantized.
The DCT algori...
This paper concerns the compression of image data. A method is proposed to detect the edge components from images and to approximate the remaining planec omponents by polynomials. The coefficients of the polynomial is vector quantized.
The DCT algorithm with average compression ratios of 10 is chosen to compare the method with existing techniques with similar compression ratios. The images compressed by DCT are rather coarse and show some blocking phenomena, which is likely due to the fact that DCT uses a basic algorithm without any classifiers. But the method proposed in this paper gives rise to fine images in which the edges are distinctly identified. In particular the whole image becomes finer when the detected edges are continuous and have no other edges around them.