Chest Radiographs Enhancement with Contrast Limited Adaptive Histogram

Hyeri Lee, Hyunseung Choo, Duc Tai Le, Junghyun Bum

Abstract: Among all the image contrast enhancement methods, histogram equalization is the most widely used. However, conventional histogram equalization causes the entire image to be too bright, resulting in noise generation as a consequence of excessive brightness changes and information loss when the histogram is intensively distributed into a specific amount of brightness. The Contrast Limited Adaptive Histogram Equalization (CLAHE) algorithm is designed to address these limitations. To avoid overamplifying brightness variations, CLAHE clips the histogram of an input image to make it uniformly distributed. This study employs the CLAHE algorithm to improve contrast in chest radiographs, with the expected result that lung disease diagnosis will be more accurate.