Video retrieval: An accurate approach based on Kirsch descriptor

In this paper, a video retrieval model is developed based on Kirsch local descriptor. In the first stage, the input video is segmented into shots and keyframes are extracted. In the next stage, local descriptors are extracted from each keyframe and clustered into k clusters using k-means clustering procedure. Given a query frame, the local descriptors are extracted from it in a similar manner, and then compared with the descriptors of the database video using k-nearest neighbor search algorithm to find the matching keyframe. Experiments have been performed on the TRECVID video segments to demonstrate the performance of the proposed approach for video retrieval applications.