Deinterlacing using edge-based motion estimation
Kyeong-Hoon Jung, Chul-Woo Kim, Choong Woong Lee · 2002
I-to-P (interlaced to progressive) conversion plays a great role in standard conversion between various image formats and the motion estimation techniques are widely used to get improved performance. In this paper, we present a new I-to-P conversion scheme for deinterlacing a video sequence. We decompose the image into four classes, i.e., unchanged, appearing, disappearing and moving area and each area is deinterlaced by an appropriate method. Especially, an edge-based motion estimation algorithm is proposed and adopted for the deinterlacing of the moving area. The proposed motion estimation algorithm greatly reduces the computational load because it uses only the edge direction instead of the pixel intensity value. Computer simulation shows that our deinterlacing scheme is reasonably applicable for I-to-P conversion.