Research And Application Of Super-Resolution Reconstruction Algorithm Based On The Sparse Theory In Medical Image

With the development of modern medical science, medical imaging technology has been widely used. Medical images include CT (Computed Tomography), MRI (Magnetic Resonance Imaging) and ultrasonic technology, etc. These technologies allow doctors to fully understand the internal structure of the patients and make accurate treatment plan. It is of great diagnostic significance.However these medical imaging is not directly obtained, they are often reconstructed by algorithm using data collected by sensors which receive X wiring harness or gamma-ray or ultrasound penetrating the human body. Therefore it had two defects:One is the resolution of the medical image is affected by the radiation dose, High radiation dose getting higher resolution but producing a greater harm to human body; Another one is the resolution of medical image are closely related with medical instrument hardware and reconstruction algorithm, meaning resolution means that the high cost of hardware and software. This article go deep into the medical image super-resolution reconstruction algorithms.This article is mainly based on the sparse representation theory and morphological component analysis and do a lot of pertinent research. In recent years, the sparse representation theory is the hotspots of image denoising, inpainting and reconstruction. Because of the sparse decomposition characteristics of morphological component analysis, MCA is becoming more and more widely used. The main content of this article is:1、Firstly expound the CT image of low resolution imaging model as the theoretical basis of super-resolution reconstruction simulation in this paper. Introduce some kind of effect index of super-resolution algorithms. Expound the basic interpolation algorithm and the interpolation algorithm of advanced technology. Finally, the article expounds the theory basis and the concept and main field of sparse representation.2、Research on image super-resolution based on interpolation. This paper briefly introduces the two major techniques in the field of image decomposition:The variation theory and morphological component analysis. The article expounds the similarities and differences and advantages and disadvantages of both. In light of different forms of medical image take different interpolation algorithm, at last synthetic high resolution image. In the process of decomposition of MCA, put forward a novel method to adaptive filtering noise. Improve the efficiency of Bilateral filtering fusion interpolation algorithm for geometric structure component.3、Research on image super-resolution based on learning. On the basis of sparse dictionary learning, Morphological component is chosen as the characteristic to make full use of image texture feature and geometrical structure characteristic, which is more suitable for the sparse expression. We put forward image model correction algorithm to improve the result of super-resolution reconstruction.4、Super-resolution technology in the application of remote medical system. Design efficient index of medical image storage system. We design a reliable, flexible, compatible back-end service framework, including medical image resolution and super-resolution reconstruction of services.