Vector quantization based image compression with Kekre, Walsh and Slant Wavelet transforms in Thepade's Transform Error Vector Rotation codebooks

Abstract

The database in the form of images has increased up to huge extent. Lack of storage space and bandwidth available are the main limitations faced to manage image databases. Most of the applications required fast transmission of image files and efficient storage. Image compression gives possible solution for all these conditions. It represents an image into reduced format without degrading its quality up to certain level so that it seems same as original image. Vector quantization (VQ) is lossy but useful technique of image compression. In this paper, Thepade's Transform Error Vector Rotation (TTEVR) algorithm with Kekre, Walsh, Slant orthogonal transforms and respective Wavelet Transform of Kekre, Walsh and Slant transforms are used for codebook generation. The experimentation is conducted with codebook sizes of 256 and 512 on a testbed of 13 assorted images of size 256×256×3 and 512×512×3. Results show that the codebook of size 256 and 512 generated from TTEVR with Kekre Wavelet Transform gives better compression quality as compared to all other transforms considered here.