Image Quality Assessment

In an age where visual intelligence products are omnipresent, images hold a significant position in digital communication. Image-based applications are ubiquitous in our daily lives. The implementation in this paper enhances the distorted image with the help of a reference image using the DISTS and ADISTS algorithms, evaluating IQA metrics, and comparing the performance of both models in image enhancement tasks. The results indicate that DISTS is particularly effective at improving traditional pixel-based metrics like PSNR which is 13.15% higher than ADISTS algorithm and aids in lowering the MSE by 17.86% more compared to ADISTS, making it ideal for tasks that demand high pixel-level accuracy. On the other hand, ADISTS shines in enhancing perceptual quality metrics such as SSIM, MS-SSIM, and VIFP, by 2.46%,0.52% and 16.7% respectively compared to DISTS algorithm indicating that it is more effectively captures structural details and enhances visual quality, for the sea anemone image which was executed for 2000 iterations for the both DISTS and ADISTS algorithms.