Iterated Adaptive Entropy-Clip Limit Histogram Equalization for Poor Contrast Images

Poor contrast and hidden details in images owing to low camera quality, bad illumination, and poor setting for environment capture are the main challenges in the contrast enhancement process. Histogram equalization (HE) based methods are commonly used to reduce these problems. However, resultant images produced by existing methods often look unnatural, are affected by unwanted artifacts, and suffer from washed-out effects. Therefore, this study proposes a local type of HE based contrast enhancement method called iterated adaptive entropy-clip limit histogram equalization (IAECHE). First, IAECHE divides the image into multiple sub-images to fine-tune the local details. Each sub-image will be applied with a different number of enhancement iterations, which are adaptively determined by IAECHE. This process ensures the best enhanced resultant sub-images, which are then merged to produce the final resultant image. A quantitative analysis of 820 images from three databases shows that IAECHE successfully produces better resultant images with better contrast, clearer details and less affected by noises. The proposed method also successfully preserves the details and structures of the images compared to several state-of-the-art methods. These findings are supported and validated by quantitative analysis. IAECHE produces high average values of peak signal-to-noise ratio, discrete entropy, absolute mean brightness error, structure similarity index, and contrast improvement index. From the analyses, the proposed IAECHE technique is proven to be capable of enhancing images' contrast, details, and structure. Thus, it has high potential for applications in the machine vision, industrial, and medical imaging fields.