Title :
Coding of image sequences using coarse quantization and feature based hierarchical block matching
Author :
Lai, M.M.L. ; Kingsbury, N.G.
Author_Institution :
Cambridge Univ., UK
Abstract :
Today, there are many different algorithms for compression of images and image sequences. They can be categorized into different types according to Netravali and Limb (1980). One of the coding methods under the group `other methods´ is contour coding methods. Contour coding methods, or coarse quantization methods, for image coding have been investigated since the early 1960´s. In 1961, Gabor and Hill proposed a one dimensional contour coding method for encoding television signals. In 1967, Graham introduced a two-dimensional contour coding scheme. In the paper, a three-dimensional contour coding scheme is proposed, with the third dimension being the temporal dimension. A feature based motion compensation technique is used to exploit the temporal redundancies of image sequences. This coding scheme is directed at transmission rates of 64 kbits per second or lower. The artifacts introduced by this coding scheme are significantly different from those of typical DCT or subband coding schemes
Keywords :
encoding; picture processing; coarse quantization; contour coding scheme; hierarchical block matching; image sequences; motion compensation technique; temporal dimension; three-dimensional contour coding scheme;
Conference_Titel :
Image Processing and its Applications, 1992., International Conference on
Conference_Location :
Maastricht
Print_ISBN :
0-85296-543-5