Hybrid feature-based transform coding of grey-level images

Abstract

The objective of image compression is to achieve the minimum possible bit-rate, within certain bounds of computational complexity, and with a tolerable degree of distortion. In very low bit-rate coding it is generally accepted that the distortions introduced will be visible to some degree. The most widely used image coding techniques today QPEG, MPEG-2,H.263), are based on block transform coding with the discrete cosine transform (DCT).