A BPNN based two-step image super-resolution reconstruction method

This paper first proposes a simple and effective nonuniform interpolation method and a deblurring method based on back propagation neural networks (BPNN). The proposed non-uniform interpolation method and deblurring method are then coupled to constitute a novel two-step super-resolution algorithm. The simulated results indicate that the proposed two-step super-resolution method shows better results than classic two-step super-resolution method. Because the nonuniform interpolation method is added before the proposed BPNN based deblurring method is performed, the BPNN is expanded to be used in uncontrolled microscanning which has non-uniform shifts between frames.