This technology is a deinterlacing method and apparatus that calculates pixel values for missing lines by fitting an approximation function, constructed via Taylor series expansion, to existing pixels in an interlaced 2D image and substituting the coordinates of the missing locations into that function. The approximation function is configured as a polynomial regression combining two 1D Taylor series expansions.
Broadcast systems such as NTSC, PAL, and SECAM have used interlaced formats to save bandwidth, but this has caused line flickering, aliasing, and line crawling in high-frequency horizontal objects. Existing intra-field interpolation techniques, which simply average surrounding pixels or consider only limited edge directions, have struggled to meet the image quality requirements of high-resolution displays like HDTVs and LCDs.
By approximating the entire neighborhood around an interpolation point as a single continuous function to estimate missing data, this technology improves both PSNR and visual quality with very short CPU processing times. It can be applied to format conversion modules in digital TV set-top boxes and broadcast transmission equipment, enabling high-quality progressive scan video even on low-spec hardware.
N/A