Application of image contours to three aspects of image processing: compression, shape recognition and stereopsis

Proposes the use of contours of constant intensity as image primitives for the purpose of image compression, object recognition and stereo matching. Compression ratios of 10-15% have been achieved, which approaches those of conventional coding techniques. The advantages of storing the images in this form is that explicit shape information contained in the image is available directly from the coded form. The contour set is partitioned into two parts, the long closed contours which tend to represent object boundaries and the shorter unclosed segments representing background detail. The former may be used as input to the object recognition and stereo matching schemes which have been demonstrated in this paper. Although further work is required before the compression ratios achieved supercede existing techniques, it is believed that the slight loss in compression ratio is a small price to concede for obtaining a compressed form which may be used directly