Optimization And Realization Of Dynamic Dimming Technology For HDR Displays System

High dynamic range displays enable viewers to experience the highest image quality,including higher contrast,deeper black states,higher gray levels,and more vibrant colors,which are one of the key technologies for next-generation displays.Due to the increased use of high dynamic range displays,reducing power consumption and improving display quality have become a major research hotspot for high dynamic range displays.In order to solve the problems of insufficient contrast and poor black state display,dynamic dimming technology has emerged at the historic moment and will become darker.Dim the display area without dimming the parts that need to be lightened,so as to achieve the purposes of increasing contrast,reducing energy consumption,improving display image quality,and so on.In order to solve the shortcomings of the traditional LCD in dynamic range and picture level performance,this paper proposes a new dynamic dimming algorithm by studying the existing high dynamic range display technology.Dynamic dimming algorithms based on image features and logarithmic functions mainly include three parts: backlight brightness extraction,backlight simulation diffusion and liquid crystal pixel compensation.At the same time,the convolutional neural network method is used to study the backlight simulation diffusion process,and the network structure and loss function are designed.Specific research work is mainly reflected in the following aspects:(1)Comprehensive analysis of the existing backlight brightness extraction algorithms,and proposed a backlight brightness extraction algorithm based on the dynamic threshold and maximum characteristics of the image area,to meet the actual brightness requirements of each image area,while maintaining the display quality while reducing the energy Consuming.(2)Compared with the existing backlight simulation diffusion algorithm,the fuzzy-diffusion method is improved,a new partition diffusion template is established,and the diffusion template coefficient is tested based on the existing HDR display system.Experiments show that the improved backlight simulation diffusion algorithm smoothes the edge backlight brightness signal better,and the computational complexity is greatly reduced.(3)This paper proposes a logarithmic function-based liquid crystal pixel compensation algorithm that complies with human eye signal processing.It stretches the image contrast and weakens the truncation phenomenon that occurs during pixel compensation.(4)Based on an independently developed HDR display system,a subjective and objective evaluation system is established to evaluate the performance of the dynamic dimming algorithm.(5)Analyzing the residual network and its application in the field of computer vision,where super-resolution is the realization of low-resolution to high-resolution,similar to the simulated backlight diffusion process in a dynamic dimming algorithm.This paper establishes a linear liquid crystal display model and proposes a backlight diffusion network based on the residual block structure,which realizes a high-precision backlight brightness diffusion process,solves the problem of obtaining complex and unreliable parameters of traditional backlight diffusion template parameters.