Image super-resolution (SR) reconstruction is to reconstruct a high-resolution\n(HR) image from one or a series of low-resolution (LR) images in the same\nscene with a certain amount of prior knowledge. Learning based algorithm is\nan effective one in image super-resolution reconstruction algorithm. The core\nidea of the algorithm is to use the training examples of image to increase the\nhigh frequency information of the test image to achieve the purpose of image\nsuper-resolution reconstruction. This paper presents a novel algorithm for\nimage super resolution based on morphological component analysis (MCA)\nand dictionary learning. The MCA decomposition based SR algorithm utilizes\nMCA to decompose an image into the texture part and the structure part and\nonly takes the texture part to train the dictionary. The reconstruction of the\ntexture part is based on sparse representation, while that of the structure part\nis based on more faster method, the bicubic interpolation. The proposed method\nimproves the robustness of the image, while for different characteristics\nof textures and structure parts, using a different reconstruction algorithm,\nbetter preserves image details, improve the quality of the reconstructed image.
Loading....