Image super-resolution reconstruction algorithm based on clustering

In view of the single frame image super-resolution reconstruction, this paper combined with sparse representation algorithm is proposed based on clustering image super-resolution reconstruction algorithm. First to sample the input image classification, clustering and for each class of training samples accordingly subdictionaries training, learning, with high and low resolution of the dictionary. Finally using the high resolution image block of dictionary and the product of the sparse representation to the low resolution image reconstruction, the experimental results show that this algorithm can effectively improve the quality of reconstruction image. In this article, through the simulation experiment and compares the traditional interpolation method, Elad method, verified the validity of the algorithm is proposed in this paper.