Abstract: An improved gamma correction image enhancement method is proposed to solve the shortcomings of unclear outline, fuzzy details and excessive noise of some MRI. Firstly, the image is transformed into gradient domain to obtain the gradient field of the image. Then, the image is divided into two different gradient intervals according to the threshold, and the corresponding gamma transform is carried out in different gradient ranges to establish the target gradient field. Finally, the enhanced image is reconstructed according to the transformed target gradient field. In the process of image reconstruction, the enhanced image is obtained by numerically solving the Poisson equation. In order to solve the problem of large amount of computation of traditional algorithms, this paper first divides the image into blocks, and then combines and stitches it. The results show that this algorithm can greatly improve the information entropy and image average difference of the original image. Taking a 512×512 MRI selected in this paper as an example, this algorithm improves the information entropy of the original image from 5.50 to 6.94, and the image average difference from 33.70 to 67.44. Experimental results show that this method can effectively improve the detail and contour information of MRI without improving the image noise.

Key words: magnetic resonance imaging (MRI), image enhancement, gamma correction, gradient field, partial differential equation (PDE), Poisson equation