Sparse Reconstruction of Magnetic Resonance Images based on Dictionary Learning: Preparation of Camera-Ready Contributions to SCITEPRESS Proceedings

Abstract

Abstract: Magnetic resonance imaging (MRI) has many advantages, such as no radiation damage to human body, high resolution and many imaging parameters. Therefore, magnetic resonance imaging technology is of great significance in the field of clinical medicine. Especially in brain imaging, because brain imaging requires high non-destructive, as long as there is a little difference in imaging results, it will cause errors in medical diagnosis. The sparse reconstruction process of MRI is based on the over complete dictionary, which has certain sparse characteristics. The main content of this paper is to K-SVD dictionary learning algorithm as the core framework, and in the process of MRI reconstruction using this algorithm. We regard dictionary learning as an equivalent process to find the optimal solution, and the learned dictionary has the core characteristics of the original data. At this time, the dictionary we get can obtain the optimal sparse representation, and reconstruct the original image through the algorithm. The experimental results show that our reconstruction results can well preserve the details of the image and obtain relatively high PSNR value.