A Classification-Based Approach for Retake and Scene Detection in Rushes Video

Retake detection has been a challenging problem in rushes video summarization. Previous approaches represent video segments as a sequence of labels then find retakes by grouping similar sub-sequences using some sequence alignment algorithm. However, these kinds of representation usually lead to unsatisfactory results because it is difficult to know the number of labels needed for a video. In our method, instead of quantizing each video segment into a label, we formulate it as a binary classification problem between pairs of segments. We use this information as the input for the Smith-Waterman algorithm to detect and group similar video sub-sequences to find retakes. Our experiments evaluated on the standard benchmark dataset of TRECVID BBC Rushes 2007 show the effectiveness of the proposed method.