Contrast enhancement of computed tomography images by adaptive histogram equalization-application for improved ischemic stroke detection

Abstract

The visualization of computed tomography brain images is basically done by performing the window setting, which stretches an image from the Digital Imaging and Communications in Medicine format into the standard grayscale format. However, the standard window setting does not provide a good contrast to highlight the hypodense area for the detection of ischemic stroke. While the conventional histogram equalization and other proposed enhanced schemes insufficiently enhance the image contrast, they also may introduce unwanted artifacts on the so-called “enhanced image.” In this article, a new adaptive method is proposed to excellently improve the image contrast without causing any unwanted defects. The method first decomposed an image into equal-sized nonoverlapped sub-blocks. After that, the distribution of the extreme levels in the histogram for a sub-block is eliminated. The eliminated distribution pixels are then equally redistributed to the other grey levels with threshold limitation. Finally, the grey level reallocation function is defined. The bilinear interpolation is used to estimate the best value for each pixel in the images to remove the potential blocking effect.