Enhancement of Brain Images Based on Nonsubsampled Shearlet Transform

In this paper, an improved brain image enhancement method based on nonsubsampled shearlet transform (NSST) is proposed. The brain image enhancement can enhance the quality of the image so that more detailed information can be seen in the brain image. Firstly, the input image is decomposed into one low-frequency subb-and and several high-frequency sub-bands with NSST: Secondly, the gradient domain guided filter is performed on the low-frequency sub-band to enhance the contrast of the image, and the improved fuzzy contrast is used to suppress the noise and adjust the coefficients of high-frequency sub-bands; Thirdly, the inverse nonsubsampled shearlet transform (INSST) is applied to reconstruct the adjusted coefficients, and the enhanced image is obtained. Experimental results demonstrate that the proposed technique can perform well in contrast enhancement, and outperforms the state-of-the art schemes in terms of the objective evaluation indexes including the H, RMSE, PSNR, SSIM, MAE, ALC and UIQI.