Backlight and Spotlight Image Enhancement Based on Von Kries Model

Backlight and spotlight pictures are images containing both dark and bright regions, caused by a non-uniform illumination. Improving the visual quality of such pictures is a challenging task since one has to increase the visibility of content and details in the dark areas without over-enhancing the bright areas. This work presents REK, a new, computationally efficient method for enhancing this kind of images. REK linearly up-scales the image channel intensities by a von Kries map and sums up the obtained image to the input one by penalizing the contribution of the up-scaled bright regions and awarding that of the up-scaled dark regions. In this way, REK improves the dark areas and preserve the bright ones. The experiments, carried out on data made freely available by the author, show that REK effectively improves the quality of backlight and spotlight images and also outperforms other state-of-the-art methods.