Implementing Histogram Equalization and Retinex Algorithms for Image Contrast Enhancement
M. Prabhu, V. Sarumathi, C. Vaishnavi · 2014
The image enhancement targets on transforming input image as better one, so that the enhanced image solve the purpose of specific application or set of objectives. Histogram equalization is a primitive and well established technique for enhancing image contrast. The existing methods are Range Limited Bi-Histogram Equalization (RLBHE) and Adaptive histogram equalization method. RLBHE works by dividing the input histogram into two independent sub-images with reference to threshold. This minimizes the intra-class variance thereby support fine separation of objects from its background. In Adaptive histogram equalization (AHE), the input source image is divided by equal number of rows and columns and the histogram for each region is calculated. The histogram of each sub region is used to perform local image enhancement. But both the RLBHE and AHE are very slow and do over enhancement of noise in the image. On the other hand, Retinex algorithms improve the sharpness, contrast and brightness of an image. It also reduces the noise level in the image. It uses OTSU-S method to threshold the images. This paper explains a walkthrough on both histogram equalization and Retinex algorithm techniques by implementing them for image contrast enhancement.