Enhanced Image Compression Via an Optimized Wavelet Thresholding Approach

Abstract

Image compression is crucial for reducing the size of digital images, particularly in applications such as medical imaging and satellite communication. Discrete Wavelet Transform (DWT) has emerged as a powerful tool for image compression due to its multiresolution analysis capabilities. In this study, we propose a novel thresholding technique for selecting optimal threshold values to enhance image compression using DWT. We evaluate the performance of our approach using different wavelets, including Haar, db2, db5, sym2, and coif1, and analyze key metrics such as Peak Signal-to-Noise Ratio (PSNR), Compression Score (CS), and L₂-norm recovery. Here, we show that for first level of decomposition using db2 wavelet, our proposed technique achieves a PSNR of 35.79 dB, a CSof 75%, and an L2-norm recovery of 99.946% when applied to a standard 474 × 474 image of Srinivasa Ramanujan. Our method not only maintains image quality but reduces computational time compared to existing thresholding techniques. This work contributes to the advancement of efficient image compression techniques in various domains.

Keywords: Image Compression; Discrete Wavelet Transform (DWT); Thresholding; Multiresolution Analysis (MRA).

2010 Mathematics Subject Classification. 41A30; 42C15; 42C40