Robust Fuzzy Bilateral Filtering Technique and Applications for Video Deinterlacing

High-Definition Intra-field Video Deinterlacing Based on Fuzzy Metric Bilateral Filtering

This technology is a deinterlacing method that calculates a fuzzy metric by applying T-norm operations to an interlaced signal, then determines the final interpolated value by processing neighboring pixels around missing pixels using a domain and range Gaussian bilateral filter.

Interlaced video often suffers from visual artifacts such as flickering horizontal lines or blurring. While motion-compensated inter-field deinterlacing offers high quality, it requires motion estimation engines and additional hardware, leading to high computational costs, whereas simple intra-field methods have limitations in interpolation quality.

This technology determines weights by reflecting the activity and directionality between the target pixel and its neighbors through a fuzzy metric, achieving edge-preserving interpolation without the need for motion estimation. It improves visual quality, PSNR, and video continuity compared to existing methods. It can be applied to TV and display image processing chips and broadcast format conversion equipment, allowing for clear playback of legacy interlaced content without hardware overhead.

‍

‍

Key Features:

• A step of calculating a fuzzy metric between pixels by applying T-norm selection to an interlaced input video signal
• A step of performing bilateral filtering on neighboring pixels of a central missing pixel using the fuzzy metric with domain and range Gaussian filters
• A step of determining the final interpolated value of the missing pixel based on the bilateral filtering results
• A step of defining the fuzzy metric space to satisfy positivity, identity, and symmetry requirements using one of the minimum, product, or Lukasiewicz T-norms

‍