Please use this identifier to cite or link to this item:
|Title:||IMAGE COMPRESSION USING WAVELET TRANSFORM AND VECTOR QUANTIZATION|
|Authors:||Rao, Chaganti Srinivas|
|Keywords:||ELECTRONICS AND COMPUTER ENGINEERING;IMAGE COMPRESSION;WAVELET TRANSFORM;VECTOR QUANTIZATION|
|Abstract:||Image compression is now essential for applications such as transmission and storage in databases. A fundamental goal of image compression is to reduce the bit -rate for transmission or storage while maintaining an acceptable image quality. Compression can be achieved by transforming the data, projecting it on a basis functions, and encoding this transform. In the present dissertation a scheme for image compression based on wavelet transform has been implemented. First, wavelet transform is used in order to obtain a set of biorthogonal sub-classes of images. The multiresolution nature of the discrete wavelet transform is proven as a powerful tool to represent images decomposed along the horizontal and vertical directions using pyramidal multiresolution scheme, result in subimages with different resolutions corresponding to different frequency bands. Second, the resulted coefficients are vector quantized (VQ) using LGB algorithm. By using an error correction method that approximates the reconstructed quantization error, distortion is minimized for a given compression rate. Finally, a lossless compression technique, Huffman coding, is applied on vector quantized wavelet coefficients to achieve better compression rate. Several 256 x 256 black-and-white images are trained together and common table codes (codebook) created. Using these tables, images inside and outside the training set are compressed and reconstructed. The reconstructed images are enhanced using median filtering method. Findings suggest that the compression ratio and the quality of the reconstructed images outside the training set have almost similar results as for the images taken from the training set. Compression ratio of 45-53, and PSNR of 31-38 dB without using Error Correction and 35-42 dB using Error Correction method have been achieved.|
|Research Supervisor/ Guide:||Joshi, R. C.|
|Appears in Collections:||MASTERS' THESES (E & C)|
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.