A Novel Shot Boundary Detection Method Based on Genetic Algorithm-Support Vector Machine

Shot boundary detection (SBD) plays an important role in content-based video retrieval. In this paper, a novel algorithm for SBD based on support vector machine (SVM) and genetic algorithm (GA) is proposed. First of all, features of pixel domain and compressed domain are synthetically extracted, and then organized into a multi-dimension vector by using the method of sliding window. Following that, the genetic algorithm is utilized to implement the simulation and iterative optimization towards parameters of SVM kernel function, then the model trained by the approximately optimal parameters is applied to judge and classify the frames of video, thus SBD is completed. The proposed algorithm solves the difficulty in parameter selection of SVM, and experimental results on the TREC-2001 video data set indicate the effectiveness and robustness of our algorithm.