De-interlacing of video images using a shortest path technique

This paper presents a fast algorithm for de-interlacing of video images using a shortest path technique. The algorithm applies dynamic programming techniques to find a shortest path in a cost matrix. The motion information obtained from this shortest path is used to re-align the fields of a video image. By using the shortest path via dynamic programming, the motion information estimated is more reliable than simply performing a search in a local neighbourhood. A variety of real images have been tested, and good results have been obtained. Keywords—De-interlacing; Shortest path; Dynamic programming; Image matching; Image motion.