Adaptive Image Compression Based on Region of Interest for Bandwidth-Constrained Environments

In the current digital era, the transmission and storage of image data in bandwidth-constrained environments pose significant challenges. Traditional image compression methods applying uniform compression levels across the entire image often fail to achieve optimal performance. This study proposes a novel adaptive image compression method that automatically detects and distinguishes regions of interest (ROI) in the image, aiming to optimize the compression ratio while maintaining image quality in critical areas. We developed an algorithm that combines Canny edge detection and Sobel gradient analysis to identify ROIs. The method was evaluated on a dataset of five test images using standard metrics: compression ratio, PSNR, and SSIM. The proposed method achieved an average compression ratio of 39.64:1, an average PSNR of 31.82 dB, and an SSIM of 0.8594. Processing time was optimized with 144.2 ms for compression and 42.9 ms for decompression. The results demonstrate the effectiveness of the proposed method in balancing compression efficiency and image quality, especially suited for constrained transmission environments.