A Local Dimming Algorithm Based on Deep Learning

In order to enhance the visual quality of Liquid Crystal Display (LCD-LED) while minimizing power consumption, local dimming algorithms have been proposed. The algorithm consists of a backlight signal extraction algorithm and a pixel compensation algorithm. It adjusts the brightness of the backlight matrix based on the characteristics of the input image, and then the pixel compensation algorithm compensates for the pixel detail loss caused by backlight dimming, thereby achieving the goal of improving contrast and suppressing power consumption. Since images with different characteristics require different local dimming algorithms, a single set of algorithms is difficult to meet the needs of different images. Therefore, this paper presents a pixel compensation algorithm based on deep learning. Empirical results demonstrate that this algorithm outperforms traditional methods in terms of contrast, peak signal-to-noise ratio, and structural similarity.