Performance analysis for two typical spatial domain super-resolution reconstruction algorithms

Image super-resolution reconstruction can use single or multiple complementary low-resolution images to reconstruct a high-resolution image. The most widely used spatial domain algorithm can combine complex motion model with corresponding interpolation, filtering resampling and iteration to reconstruct a much more better image. Two typical super-resolution reconstruction algorithms in spatial domain are investigated in this paper: Iterative Back Projection (IBP) and Projection onto Convex Sets (POCS). They are compared through various super-resolution zoom experiments, such as traffic monitoring images, faces, chicken´s cardiovascular OCT images and other types of images. The results are analyzed and compared by visual effects, root mean square error (RMSE), the peak signal to noise ratio (PSNR), structural similarity (SSIM), carrier noise ratio (CNR), and algorithm complexity. Finally, the corresponding application situation of two algorithms is given.