Object level frame comparison for video shot detection with orthogonal polynomials model

Shot Detection (SD) is the fundamental step towards video indexing and retrieval system and hence researchers pay effort on developing accurate shot change detection methods. However, the high computation cost remains as a concern. In this paper, an Object Level Frame Comparison (OLFC) for video shot detection with orthogonal polynomials model is proposed in the compressed domain to reduce the computation cost and to increase the accuracy. The proposed method utilizes orthogonal polynomials model coefficients at a block level to represent background and adapt the background by updating the proposed orthogonal polynomials transform coefficients. The proposed object segmentation approach can extract the foreground objects with pixel accuracy through a two-stage process. First, a new background subtraction technique in the orthogonal polynomials transform domain is exploited to identify the block regions fully or partially occupied by foreground objects and then pixels from these foreground blocks are further classified in the spatial domain. The experimental results show that the proposed background modelling algorithm can achieve comparable accuracy to their counterparts in the spatial domain, and the associated segmentation scheme can visually generate good segmentation results with less computational effort. The extracted objects with the proposed orthogonal polynomials model are then subjected for Object Level Frame Comparison (OLFC) technique to detect the shot change. Experimental results reveal that the proposed system outperforms the existing systems with high precision and recall rate.