Image Compression with Edge Local Variable Based Inpainting

Abstract

Image compression is a vital process in digital image storage and transmission, aiming to reduce data size while retaining acceptable visual quality. Traditional compression techniques, however, often result in artifacts such as blurring, blocking, and loss of structural details, especially along edges which are fundamental for human perception. To overcome these shortcomings, this work explores edge local variable based inpainting as an effective strategy to enhance image compression. The proposed approach extracts edge-related variables such as orientation and gradient information, which are then utilized to guide the inpainting process for reconstructing regions affected by compression. By prioritizing the preservation of edges and structural features, the method ensures higher visual fidelity while achieving significant data reduction. Experimental results indicate that the integration of edge local variable based inpainting not only improves compression efficiency but also enhances perceptual quality, making it highly suitable for applications in medical imaging, remote sensing, and multimedia communication.