Image super-resolution reconstruction algorithm based on coordinated attention generative adversarial network

Aiming at the problem that the parameters of the existing image super-resolution reconstruction model are too large to be applied in reality, a single-image super-resolution reconstruction model——Generative Adversarial Network Based on Coordinate Attention Mechanism(CSRGAN) is proposed. By optimizing the generator in SRGAN, the coordinate attention mechanism is combined with the residual network to construct the CR module, which promotes the flow of information between channels and strengthens the feature selection ability of the network; at the same time, a hierarchical feature fusion is built in the main network structure, improve the utilization of early features in the deep network, and a large number of long and short jump connections alleviate the gradient disappearance and improve the network convergence speed. Tested with RFB-ESRGAN, ESRGAN and other models on Set5, Set14, BSD100 and Urban100 data sets, the Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity (SSIM) have been improved, and the number of model parameters has been greatly reduced, the reconstructed image has improved clarity and structural integrity.