Entropy Based Model For Lossy Image Compression Scheme Using Wavelets, Svd, And Two-Channel Coding Techniques

Abstract

This study introduces a novel image compression technique that leverages wavelet transforms, Singular Value Decomposition (SVD), and two-channel coding methods to achieve high compression ratios while maintaining perceptible image quality. Wavelet transforms offer a multiresolution image representation, preserving essential information while compressing less critical features. SVD further reduces dimensionality by decomposing the image into singular vectors and values, allowing for a customizable balance between image quality and compression ratio. The two-channel coding techniques enhance compression efficiency by separating image statistics into two channels—one dedicated to storing crucial image data and the other to encoding supplementary information. The entropy-based model dynamically allocates bits to each channel, prioritizing the most salient image properties during compression. The method was evaluated using both qualitative and quantitative metrics, including Bits Per Pixel (BPP), Structural Similarity Index (SSIM), and Peak Signal-to-Noise Ratio (PSNR), on benchmark grayscale image datasets. Experimental results demonstrate that the proposed compression scheme outperforms existing methods, showing significant improvements in compression ratio, saving percentage, and bits per pixel, with averages of 40.72%, 31.79%, and 69.35%, respectively, compared to the JPEG method. Additionally, comparative analysis using PSNR, SSIM, and entropy metrics highlights the scheme's superior performance in quality enhancement and visual quality preservation over JPEG. This approach holds promise for applications in clinical imaging, remote sensing, and multimedia communication, contributing to the advancement of lossy image compression techniques and addressing the growing need for efficient compression in a digital-centric world.