Shot boundary detection is a major and challenging job in any video processing solutions. Different features are used for detecting shot boundaries in a video. Choice of features depends on the nature of processing. Mostly features are extracted from either spatial domain (visual features) or from temporal domain (motion information). To obtain more accuracy normally high level features are used but these features need complex computations. On the other hand low-level features like colour and brightness are easy to extract and compute but these features have less tolerance towards small changes in the frame. Similarly features in temporal domain have their limitations — they mostly ignore the spatial domain and hence give false results. In this proposed solution, features from both spatial and temporal domains are combined. In spatial domain brightness information is used that is easy to extract and fast to compute. From spatial domain motion information is used. Motion information is extracted from sub-blocks that add spatial information as well. Finally both the results are combined resulting in higher accuracy, more sensitivity and tolerance.
Weitere Kapitel dieses Buchs durch Wischen aufrufen
Bitte loggen Sie sich ein, um Zugang zu diesem Inhalt zu erhalten
Sie möchten Zugang zu diesem Inhalt erhalten? Dann informieren Sie sich jetzt über unsere Produkte:
- Short Boundary Detection Using Spatial-Temporal Features
Neuer Inhalt/© ITandMEDIA, Product Lifecycle Management/© Eisenhans | vege | Fotolia