A New Compression Method of Gray Image

Using similarity among local images and geometric moments, a new gray image compression method based on geometric moment invariants has been proposed. The original image is divided into non-overlapping sub-images, and the geometric moments and geometric moment invariant of each subimage are computed. If the sub-images in neighborhood have the same moment invariant, the latter sub-image is discarded with additional information recorded. Considering the size of subimages is small, every sub-image is replaced by its average brightness to obtain more compression ratio, and cubic spline interpolation is adopted in decompression. The proposed method is compared with DCT algorithm, and good visual effect has achieved.