An Edge-Guided Exact Histogram Specification Method for Enhancing Extremely Dark Images

Abstract

Histogram equalization (HE), and especially exact histogram specification (EHS), is a commonly used method for enhancing images. Although the EHS makes it possible to transform an image to follow exactly any specified target histogram through an image intensity mapping function, how to practically choose an appropriate target histogram remains an open problem. This paper first shows that, although the goal of image contrast enhancement focuses mainly on the image edges, the EHS performance is also affected by the image non-edge content. Then it proposes a new edge-guided EHS method that solves such drawbacks. In this method, it first carries out the exact histogram equalization (EHE) on only the edge points, and then it applies the resulting intensity mapping function to the whole image. This generally results in a non-uniform histogram for the transformed image. Experimental results show that the proposed method is very efficient and achieves state-of-the-art performance for enhancing extremely dark images.