Super Resolution of Images Using Residual Network
Single Image Super-Resolution using residual learning for high-quality image reconstruction.
Project Overview
This project explores Single Image Super-Resolution (SISR) using a Residual Network (ResNet) to reconstruct high-resolution images from low-resolution inputs.
The model uses residual learning to preserve image structure and fine details and was trained on the DIV2K dataset. Reconstruction quality was evaluated using PSNR and SSIM.
Results
The model achieved a PSNR of 30.25 dB and an SSIM of 0.77 on the DIV2K dataset, demonstrating its ability to reconstruct higher-resolution images while preserving structural information and visual detail.
Publication
Super Resolution of Images Using Residual Network
Published in the Proceedings of the 3rd International Conference on Futuristic Technology (INCOFT 2025).