A novel image super-resolution reconstruction method based on sparse representation using classified dictionaries

This paper presents a new approach to single-image super-resolution reconstruction, based on sparse signal representation using classified dictionaries. The high-resolution and low-resolution image patches training sets are divided into two categories respectively by two new classification templates which give consideration to direction and edge features. Then, we train a pair of learning dictionaries which illustrate the features of direction and edge by learning both kinds of image patches. Learning dictionaries combined with sparse signal representation to realize the image super-resolution reconstruction. As the experiment results show, the new method has good perforins to restore the lost high-frequency information, and its robustness is good.