DENOISING AND SUPER-RESOLUTION OF MEDICAL IMAGES BY SPARSE WEIGHT METHOD

Abstract

High resolution images are needed for many applications like medical imaging for diagnosis and treatment. Denoising and super-resolution of medical images is proposed in this paper. Construct a database of high and low resolution image patch pairs and with the help of this database estimate a high resolution image from a single noisy low resolution image. In order to find out a high resolution version from the given input low resolution version, get the non-negative sparse linear representation of the input patch over the low resolution patches from the database. Non-negative quadratic programming approach is used for the sparse process. For the low resolution and noisy images, it is a widely adequate method. Edges of the super resolved image can be enhanced by using the blind deconvolution algorithm.