A locally-adaptive approach for image gamma correction

This paper proposes a new adaptive gamma correction method. The method estimates gamma values in different regions of an image and transforms individual pixels of the image following estimated gamma values. In this method we initially made a database of two feature vectors: pixel intensity histograms and dispersion vs. location distributions, from nine standard images with different gamma conditions. After computing these feature vectors for each region of the input image, the most appropriate gamma value is found from the database using a K-Nearest Neighbor technique. One of the advantages of the proposed method is that the gamma value for each pixel of image is modified separately. Experimental results demonstrate that the proposed method performs better in estimating gamma values and improving image quality.