Low-Level Vision Based Super-Resolution Image Reconstruction
Kai Xie, Haixia Guo, Guo Hai-long · 2007
In the process of super-resolution image reconstruction, corner detection and interpolation are two key technologies. In this paper, we proposed two improved methods for them. Firstly we propose a variable threshold for the allowed variation in brightness within the USAN area. The approach makes corner well-distributed and can reduce lost and false corners relatively. Experiments confirm that the method is anti-noisy and has the less computation. Secondly a new adaptive interpolation approach based on circular-area is presented. The approach can adoptively select the interpolation method based on the gray feature of an image.