Bio-Medical Image Enhancement Using Adaptive Multi-Resolution Technique

Abstract

The enhancement, resolution, and accuracy are the basic requirements of digital image processing. But in the field of medical, the local details of the images are necessary. In this contest, the Contrast Limited Adaptive Histogram Equalization (CLAHE) is an effective approach, which is used for enhancing the local feature. The downside of CLAHE is the contrast overstretching and noise enhancement. To overcome this problem the low frequency and high-frequency details of an image were considered separately. This paper proposes a new biomedical image enhancement technique based on Bi-dimensional Empirical Mode Decomposition (BEMD) with CLAHE and Sobel Gradient. Where the image is decomposed using BEMD to separate the edge and detail information. The edge information is enhanced using CLAHE and the rest coefficients remain unchanged. Experiments are conducted on numerous brain, retina and tissue images to compare the performance of the proposed algorithm with the state of art approaches. The comparison is made quantitatively in terms of entropy ratio (ER), absolute mean brightness error (AMBE), the peak of signal to noise ratio (PSNR), structural similarity index (SSIM), and image quality index (IQI); qualitatively through visual comparisons. The results indicate a significant improvement especially in terms of structural and edge preservation.