A Superresolution Algorithm Based on Image Color-Texture Feature Classified Learning

Abstract

Example-based superresolution algorithms obtain priori knowledge to take part in the majorization of reconstruction procedure by machine learning. However, the wrong matching results during the sample database searching lead to image degradation, which is difficult to ensure the overall quality continuity of restored images. In this paper, we propose a new superresolution algorithm based on image color-texture feature classified learning. The algorithm saves the corresponding color-texture information into the sample set and selects object subset by Support Vector Machine (SVM) pre-classified learning. Then in the high frequency prediction process we make precise matching search from the subset of sample database which has similar color and texture features with the object image. Contrast experiments show that the proposed optimization algorithm removes the samples which have little content relevance, shortens the program running time, reduces the mis-matching greatly and improves the image rebuilding performance.