Content-based video shot boundary detection using multiple haar transform features

Abstract

In the entire content-based video technologies, video shot boundary detection (VSBD) is a crucial task. VSBD is used to segment the video into shots. In this paper, we propose a VSBD technique which identifies the abrupt (cut) gradual transitions in video sequence. In our technique, we first extract the multiple feature vectors from video sequence using discrete Haar transform (DHT) for N = 4. The similarity between successive video frames is measured using extracted features, and a continuous signal is computed. To identify shot transition, procedure-based shot detection algorithm is applied on continuous signal. The proposed technique is evaluated on TRECVID test collection videos. The experimental results show that the proposed technique performs better than the existing method in identifying cuts and gradual transitions.