Image Super-Resolution Reconstruction by Blur Kernel Estimation

The paper combines blur kernel estimation and dictionary learning and proposes an image super-resolution reconstruction algorithm based on Blur Kernel Estimation (Dictionary Learning based on Blur Kernel Estimation, DL-BKE). The proposed algorithm performs blur kernel estimation on the test image before reconstruction of the test image and replaces the blurring process in image preprocessing with the obtained blur kernel, and then reconstructs the high-resolution image using the dictionary learning method. The experimental results show that the DL-BKE algorithm robustly reconstructs better high-resolution images for test images with different blurring degradation processes compared to the classical super-resolution algorithm.