Optimization Algorithm for Local Dimming of EdgeLit TVs towards High Contrast and Low Power Consumption

摘要

Local dimming algorithm for Edge-Lit LCD is a severe challenge, as the light of an LED string is not locally but widely distributed over the panel. A further challenge is the pixel compensation for locally dimmed backlight (BL). Avoiding clipping and at the same time preserving the image brightness is a problem. In our approach the LED backlight is modeled on a physical base by light spread function (LSF) considering the crosstalk between LEDs. Discrete optimization method is used to determine the LED values with the scope of minimizing power consumption. Image correlation method has been amended to further increase the static contrast. Temporal filter are applied allowing smooth video operation. Based on image analysis, specific constraints are added to make the system robust against production variation e.g. LED, LSF etc. This local dimming technology preserves the image brightness, avoids loss of detail and decimates the undesired "clouding" artifact of Edge-Lit TVs. High power saving is achieved, while the HW cost is low. It has been proven at a premium serial TV and automotive displays.