A Hybrid Approach to Compress Still Images using Wavelets and Vector Quantization
S. Vimala1, P. Usha Rani2, J. Anitha Joseph3
1Dr. S. Vimala, Department of Computer Science, Mother Teresa Women’s University, Kodaikanal, (Tamil Nadu), India.
2P. Usha Rani, Department of Computer Science, Mother Teresa Women’s University, Kodaikanal, (Tamil Nadu), India.
3J. Anitha Joseph, Department of Computer Science, Mother Teresa Women’s University, Kodaikanal, (Tamil Nadu), India.
Manuscript received on March 13, 2015. | Revised Manuscript received on March 26, 2015. | Manuscript published on April 30, 2015. | PP: 56-59 | Volume-4 Issue-4, April 2015. | Retrieval Number: D3852044415/2013©BEIESP
Open Access | Ethics and Policies | Cite
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: This paper presents a hybrid technique for compression using Wavelet and Vector Quantization (VQ). Wavelet is a technique for representing the image into various degrees of resolution. The input image of size 256 *256 pixels is divided into 4 sub-bands named LL, HL, LH, HH by applying Discrete Wavelet Transform. Vector Quantization is then applied for the lower sub band (LL). The size of lower sub-band is 128*128 pixels. VQ is a lossy image compression technique used to have improved coding efficiency. In the proposed study, the different types of wavelets such as Haar Wavelet, Coiflet Wavelet, Symlet Wavelet, Daubechies Wavelet and Bioorthogonal Wavelet are applied to the input images and the respective lower bands are then subjected to Vector Quantization in the Encoding process. The compressed image is then transmitted or stored in the form of Codebook and the Index Map, which are the outcomes of VQ. In the decoding phase, an image of size 128 x 128 pixels is reconstructed from the stored/transmitted Codebook and Index map. The reconstructed image is then subjected to Inverse DWT to get an output image of size 256 x 256 pixels. Standard images such as Lena, Baboon, Boats, Bridge and Cameraman are used to test the performance of the proposed method. With all the wavelets, the proposed technique leads to better compression ratio without losing the visual effect.
Keywords: Image compression, Wavelet, Vector quantization, Haar, coiflet, Symlet, Daubechies, Bioorthogonal.