Volume : V, Issue : IV, April - 2015
Classified Codebook with Indexmap Compression for Vector Quantization to Compress still Images
Dr. S. Vimala, Ms. S. Ezilarasi
Abstract :
Vector Quantization (VQ) is one of the Lossy Image Compression techniques. VQ comprises of three
steps in compressing the images: 1. Codebook Generation, 2. Image Encoding and 3. Image Decoding.
The phase of codebook generation plays a vital role in compressing the image. The performance of VQ depends on
the quality of the codebook. The quality of the reconstructed image is improved at the cost of compression rate and
vice versa. The feature of inter–pixel redundancy is exploited in this method. Codebook is a collection of blocks (vectors) selected at random from the input image blocks. In the proposed method, the vectors in the codebook are classified into shade (high–detail) and edge (low–detail) blocks. For shade blocks, only the average of the components of the
vector is stored, leading to a significant reduction in bpp (bits per pixel). Further compression is achieved by compressing the indices of the Index Map that is generated during the Encoding phase. The proposed method is tested with
some bench mark images such as Lena, Cameraman, Baboon, Boats and Bridge. Results are generated for codebooks
of sizes 128, 256, 512 and 1024. In all the cases, the proposed method outperforms the existing technique in terms of
bpp (bits per pixel).
Keywords :
Article:
Download PDF
DOI : 10.36106/ijar
Cite This Article:
Dr.S.Vimala, Ms.S.Ezilarasi Classified Codebook with Indexmap Compression for Vector Quantization to Compress still Images Indian Journal of Applied Research, Vol.5, Issue : 4 April 2015
Number of Downloads : 581
Dr.S.Vimala, Ms.S.Ezilarasi Classified Codebook with Indexmap Compression for Vector Quantization to Compress still Images Indian Journal of Applied Research, Vol.5, Issue : 4 April 2015
Our Other Journals...
-
International Journal of
Scientific Research Visit Website -
PARIPEX Indian Journal
of Research Visit Website -
Global Journal for
Research Analysis Visit Website