Adaptive edge-based image contrast enhancement using multi sub-histogram analysis

Edge-based image contrast enhancement is a contrast enhancement method that focuses on the thickness of edge. Splitting histogram of an image into two sub-histograms, which is called bi-histogram, can be used to enhance the image contrast based on edge. However, image histogram has various modality, which causes bi-histogram separation may be incorrect. This paper proposed an adaptive edge-based contrast enhancement method using multi sub-histogram analysis. Hierarchical cluster analysis (HCA) splits the histogram iteratively to build multi sub-histogram. Then, the edge-based contrast enhancement process is performed with adaptive plateau limit for generating probability and cumulative density function for each sub-histogram. Finally, the enhanced image is achieved by the transformation function with guided filter. Assessment of the proposed method for evaluation is using absolute mean brightness error (AMBE), standard deviation (STD), contrast improvement index (CII), discrete entropy (DE), and perceptual image sharpness index (PSI). The evaluation shows that the proposed method has the best AMBE, CII, and DE of 6.53, 9.49, and 6.21, respectively. It means that the proposed method can maintain the brightness of the image, change the contrast significantly, and provide better edge information extraction, respectively. Therefore, the assessment proves effectiveness of the proposed method to enhance the contrast by separating regions of histogram correspond to the number of the modal contained in the histogram.