MEMC Circuit Halo Mitigation via Cover-Uncover Area Detection

The patent divides the image processing into multiple stages: motion estimation, halo detection, and selective filtering. The halo effect is detected by comparing motion vectors with gradient information, and filtering is applied selectively only to areas where halo is detected, rather than processing the entire image uniformly. This segmentation allows frame rate enhancement while minimizing halo artifacts.

The patent applies different processing strengths to different regions of the image based on local characteristics. The filtering strength is adjusted according to the detected halo severity in each local area, preserving edge sharpness in regions without halo while reducing halo effects in affected regions. This local adaptation resolves the contradiction between overall quality improvement and localized artifact reduction.

The patent implements dynamic adjustment of filtering strength based on local halo detection results. The filter strength is not fixed but adapts to the actual halo conditions in each region, applying stronger filtering only where needed and weaker filtering elsewhere. This dynamic approach prevents the edge flicker that results from uniformly strong filtering while still reducing halo width effectively.

The patent uses a feedback mechanism where the halo detection result guides the filtering operation. By comparing motion vectors with gradient information, the system detects halo presence and adjusts filtering accordingly. This feedback loop ensures that filtering is applied appropriately to reduce halo without causing excessive edge flicker, as the filtering strength is continuously adjusted based on detected halo characteristics.