Image Enhancement by Recursive Correction of its Average Brightness
Sergei Yelmanov, Yuriy M. Romanyshyn · 2021
This article discusses improving the efficiency of image and video enhancements in real-time apps. The aim of this work is to increase the efficiency of improving image quality in automatic mode by transforming image intensity using adaptive piecewise-linear stretching. To this end, we propose two new approaches to adaptive transforming image intensity, which are based on the recursive mean-separate contrast stretching (RMSCS) and the recursive correction of average image brightness (RCAB). These approaches are based on the assumption that the best for visual perception is the normalized image in which the average brightness is equal to the middle of the brightness scale. The RMSCS and RCAB techniques provide a more effective improvement in the image quality, a more even distribution of contrast of objects in the image, increasing its overall contrast, and reducing the risk of distortions and artifacts. The RMSCS and RCAB techniques work in a fully automatic mode, are efficient, easy to implement, computationally low-cost, and are designed to normalize and enhance video in real-time apps.