Video Important Shot Detection Based on ORB Algorithm and FLANN Technique

A shot is continuous frames taken by a single camera. A shot is the essential unit in the analysis of video files for many applications, such as, online video advertising and e-commerce. Therefore, the important shot determines how much a shot influences the video analysis process. Because of their low substance, some video shots should be discarded in order to reduce video raw data using the summarization technique. A robust method for important shot detection and its key frame is proposed in this paper. Many parameters that indicate the important shot and its key frame are taken into account by the proposed strategy. These parameters (which include the shot’s interest points based on ORB(Oriented Rotated Brief Very Fast Binary Descriptor) algorithm and FLANN(Fast Library For Approximate Nearest Neighbors) technique, entropy metrics for video summary, shot length, and shot activity factors) have been used to determine shot relevance. Each of these elements has a different impact on the importance score. The robustness of the proposed method was demonstrated by the results.