This technology is a demosaicing method that restores missing colors for pixels of interest in Bayer CFA images by establishing O, Q, and S pentomino-shaped interpolation paths. It calculates missing colors using weighting factors derived from gradients along these paths, performing the process from direction estimation to color calculation in all four cardinal directions.
Digital camera demosaicing relies on interpolation to numerically estimate surrounding color planes, which often leads to artifacts such as jagged edges in restored images. Increasing computational complexity to mitigate this issue has historically resulted in slower system processing speeds.
This technology enhances edge estimation in conventional bilinear interpolation by adding pentomino path-based gradient settings and determining green, red, and blue components step-by-step based on difference values using inter-plane weights and polling maps. This increases robustness while keeping computational requirements low enough for real-time processing on existing hardware. It can be applied to image signal processing pipelines in digital still cameras and mobile devices, enabling sharp, high-quality color images without hardware upgrades.
N/A
