AnAudio-Visual Saliency ModelforMovie Summarization

A saliency-based methodforgenerating video subjective. Hence,mappinghumanperception intoan summaries ispresented, whichexploits coupled audiovisual automated abstraction process hasbecomequite common. In information frombothmediastreams. Efficient andadvancedanattempt toemulate themultimodal nature ofhuman speech andimage processing algorithms todetect keyframes that understanding, theInformedia project anditsoffsprings, areacoustically andvisually salient areused. Promising resultscombined speech, image, natural language understanding and areshownfromexperiments onamovie database. imageprocessing toautomatically index video forintelligent search andretrieval (9)(10)(11). Thisapproach generated Keywords-saliency; saliency curves; attention modeling; event interesting results. Goingonestepfurther towardshuman detection; key-frame selection; video summarization; audiovisualperception, Ma etal. proposed amethodfordetecting the