Research Of Sparse Representation Super-resolution Technology Based On Classified Dictionaries

High resolution images is not only to better meet the needs of computer applications, butit is important to be able to meet the human visual sensory needs. But due to limitations inimage acquisition devices at this stage, people are often unable to obtain high-resolutionimages quickly and easily. So for low resolution image reconstruction, improving imageresolution is important to research and practical applications.Technical difficulty of super resolution reconstruction is that the low resolution imagewas not sufficient to restore the actual observed high resolution image. Mathematically, theproblem is an ill-posed. Super-resolution reconstruction algorithm have many branches. Andsuper-resolution reconstruction algorithm based on sparse is more prominent now. Sparserepresentation super-resolution was based on theory of compressed sensing datarepresentation model, through the process of establishment of over-completed dictionaries,using samples of high resolution and reconstruction image patch similarity between thesample patch and high resolution patch, to reconstruct of target images.In this article, super resolution reconstruction technique is discussion and analysis.Research is mainly target at single image sparse representation super-resolution technique.The main contribution of this article is as follow:1. This paper describes and analyzes theory and application of Super resolutionreconstruction technique. And describes mainstream super-resolution reconstructiontechnology and analyzes the main problems in the SR reconstruction. Finally, discussed thereconstruction of the objective evaluation criteria.2. This paper analyses and discusses a sparse representation of the image andsuper-resolution reconstruction algorithm based on sparse representation. Improved the imagefeatures used by over-complete dictionaries during construction. From the experimentalresults, improved sparse representations image features of SR reconstruction algorithm hasbetter recoverability than the original algorithm in image quality.3. Improved sparse representations hyper structure dictionary in the reconstructionprocess. Improved methods of training set. Firstly, classify image block by image feature, and then clustered each type of features by taking a certain number of training features to traindictionaries. Finally merge the training result of the dictionaries for a complete dictionary. Bythis way you can avoid drawback of the original dictionary of construction method in randomfeature selection and bias issues. Improvements to the dictionary are better able to adapt todifferent textures, different characteristics of the image, making the image reconstructionwork better.