CScGAN: Leveraging Compressed Sensing and Generative Adversarial Networks for Image Super-Resolution Reconstruction

We propose a novel image super-resolution technique, the Compressive Sensing combined with Generative Adversarial Network model (CScGAN). During the image super-resolution reconstruction process, artifacts that arise from the complexity of the reconstruction network, imbalances in training data, and instability in the training process can degrade the quality of the reconstructed images. Our method concentrates on minimizing reconstruction artifacts through advanced signal processing and structural optimization of the network. We elucidate our CScGAN model using architectural diagrams and mathematical notation. When compared with multiple public datasets and state-of-the-art methods, our approach significantly reduces the impact of artifacts during the reconstruction process and enhances the quality of the reconstructed images.