Contrast Enhancement Using Sensitivity Model-Based Sigmoid Function

Seung Park,Sung-Jea Ko,Yong-Goo Shin

doi:10.1109/access.2019.2951583

Abstract

For indirect contrast enhancement, researchers have proposed various transformation functions based on histogram equalization and gamma correction. However, these transformation functions tend to result in over-enhancement artifacts such as noise amplification, mean brightness change, and detail loss. To overcome the limitations of conventional transformation functions, this paper introduces a novel sigmoid function based on the contrast sensitivity of human brightness perception. In the proposed method, the contrast sensitivity of the human retina is modeled as an exponential function of the log-intensity, and a transformation function is derived using the sensitivity model as the exponent of Steven’s power law. We also present a parameter optimization method that maintains the mean brightness of the input image and stretches the image histogram while minimizing information loss. Experimental results demonstrate that the proposed method has low computational complexity and outperforms the state-of-the-art methods in terms of contrast enhancement performance, mean brightness preservation, and detail preservation.

Highlights

Digital images often have low contrast owing to inadequate image capture devices or undesirable lighting conditions
Experimental results indicate that the proposed method outperforms the state-ofthe-art methods with regard to mean brightness preservation, detail preservation, and contrast enhancement performance
EXPERIMENTAL RESULTS 550 test images were taken from three datasets in [41]–[43] for a comparison of the proposed method with weighted adaptive histogram equalization (WAHE) [1], contextual and variational contrast (CVC) [20], layered difference representation (LDR) [11], adaptive gamma correction (AGC) [23], fuzzy-contextual contrast enhancement (FCCE) [45], two sigmoid function-based methods [46], [47], and some recent direct methods [38], [48], [49]

Summary

Introduction

Digital images often have low contrast owing to inadequate image capture devices or undesirable lighting conditions. Since low contrast images may have a washed-out appearance or do not reveal all the scene details [1], researchers have proposed various enhancement methods to improve the visual quality of these images [2], [3]. In direct methods [5]–[8], image contrast is measured based on the human visual system (HVS), such as the Weber–Fechner law or Retinex theory [9], and is improved by applying various nonlinear functions [5], [6] or solving optimization problems [7], [8]. Rods are very sensitive to light and provide achromatic vision named scotopic vision at low luminance levels (10−6 to 10cd/m2). At luminance levels between 10−2 and 10 cd/m2, both rods and cones are active and the human retina operates in a transition mode called mesopic vision. To describe the retinal response of human brightness perception, various response models based on neuroscience experiments have been proposed

Methods

Results

Conclusion