This technology is an image quality enhancement device that performs deinterlacing based on inferred rules. It utilizes a fuzzy rule knowledge base where video motion, pixel position, interpolation method, and window size are defined as fuzzy variables, converting the motion and position of the input video stream into fuzzy sets.
Criteria such as video motion speed or the uniformity of a pixel's region often have ambiguous boundaries, making them difficult to classify into clear-cut categories like "fast" or "slow." Conventional deinterlacing methods, which process this ambiguous expert knowledge using fixed criteria, have limitations in accurately selecting interpolation methods tailored to specific image characteristics.
This technology selects the optimal interpolation for each pixel using fuzzy rules: intra-field interpolation for fast motion, inter-field interpolation for slow motion, large windows with inter-field interpolation for uniform regions, and small windows with line-average interpolation for non-uniform regions. This improves both deinterlacing accuracy and processing efficiency. It can be applied to broadcast transmission equipment, interlaced video converters, and digital TV image processing chips, ensuring clean playback of interlaced sources on progressive displays.
N/A
