Poor Quality Image Enhancement using Histogram Equalization Adaptive Sigmoid Function

Abstract

In this article, we provide a method for improving low-resolution photographs using the Bi-Histogram Equalization technique. One of the most common ways to improve contrast is by using histogram equalization. However, because of the over-enhancement nature, it might result in an unnatural look. In the proposed work, the RGB model is transformed into the HSV model, where Hue (H) and Saturation (S) represent the color channel value represents the luminance intensity and the V-channel is considered for image enhancement. To determine the cumulative density function and probability for every sub-histogram, we have made use of two sigmoid functions. These functions will be equalized independently using histogram equalization and stretched to achieve a smooth curve. After applying the histogram distribution to each sub-histogram, the origins of the sub-histograms are placed on their medians. The results are computed using the proposed methodology and also compared with the existing methodologies. The proposed method is observed to provide better accuracy than the other existing methods.