Shot boundary detection based on Laplacian eigenmaps

Shot boundary detection senders as the preliminary´ step to structure the contents of videos, many research efforts have been devoted to this field. In this paper, we propose a novel approach to shot boundary detection in video sequences including cuts, fades, dissolves. Firstly, we propose an effective feature selection mechanism using modified Laplacian eigenmaps. which removes the features of raw feature space obviously affected by camera/object motion, illumination and other artifacts on the curve of frame-wise histogram difference. Secondly, we define different similarity measures according to different types of shot, in which pattern similarity can easily be detected. A good compromise between speed and accuracy is achieved. It also has good noise tolerance characteristics. The effectiveness of our methods has been validated by experiments on several real-world video sequences.