An algorithm for automated clip-limit selection in contrast limited adaptive histogram equalization using salp-swarm optimization, and visual information fidelity metric for radiographic applications

Contrast-limited adaptive histogram equalization (CLAHE) is widely used for improving the quality of radiographs. Quality of the radiographs processed by the CLAHE depends on the selection of a user-defined parameter, ‘clip-limit’ (CL). We propose a method based on metaheuristic optimization to automate the selection of the CL in CLAHE deployed for contrast enhancement applications in radiography. We use the visual information fidelity (VIF) statistics as the fitness function to measure the variation of quality of enhanced radiographs obtained from the CLAHE in response to the changes in the CL. We use salp-swarm optimization algorithm (SSOA) to locate the maximum value of the VIF statistics, and corresponding optimum CL. On 100 radiographs, the CLAHE-SSOA-VIF framework shows VIF higher than that shown by state-of-the-art methods used in the literature to improve the quality of the radiographs, namely histogram equalization (HE), unsharp masking (USM), and morphological image enhancement (MIE), and blind/reference less image spatial quality evaluator (BRISQUE) scores less than HE, USM, and MIE. The VIF higher than 1, portrays the ability of the CLAHE-SSOA-VIF framework to improve the contrast of radiographs without inducing any distortions. Lowest value of BRISQUE indicates that enhanced radiographs of CLAHE-SSOA-VIF framework are natural in appearance.