Streamlining the histogram equalization of digital images by using decimation approach

As smart-phones become more and more popular, there is growing demand on the mobile version of image processing applications. Due to the limitations on the mobile hardware, the computational load and the amount of data transfer should be more efficient. In this paper, we try to improve the implementation of the histogram equalization algorithm by leveraging the decimation approach. The conventional histogram equalization process includes two major steps. First, the original image is full scanned to find its gray-scale distribution. Then, the gray-levels are transferred so that the accumulated number of pixels is proportional to the assigned gray-level. In this paper, we apply the decimation process to the original image before performing the statistical procedure in the first step. The gray-level distribution of the reduced image is applied to determine the transfer curve of gray-level reassignment. The experimental results show that the proposed method is efficient and the difference with respect to the conventional algorithm is very small and invisible to the human eyes.