Method and device for extracting key scene in video

Abstract

The invention provides a method for extracting a key scene in a video, aims to solve the problems of semantic gap, large-scale data and rapid processing in the process for extracting the key scene inthe video. The method comprises the following steps: acquiring video stream information, and converting into a video frame set and audio signals; performing clustering to form a lens set; performing time-dividing on the audio signals according to a time index of the lens set to form lens audio segment signals; extracting an advanced energy engin values for the lens audio segment signals; determining composite advanced engin values according to the advanced energy engin values and the change rate of the advanced energy engin values; sorting the composite advanced engin values corresponding to the lens set to obtain a key shot set; performing expanding to form a key scene set; and determining key scenes in the video according to emotional richness and advanced energy engin values of the keyscene set. The invention also provides a corresponding apparatus, a device, and a media. Large-scale video data can be rapidly processed, and key scenes can be acquired.

Description

technical field [0001] The invention belongs to the technical field of video processing, and in particular relates to a video key scene extraction method, an extraction device, a computer-readable medium and electronic equipment. Background technique [0002] In recent years, with the rapid development of network technology and multimedia technology, video application websites have been quickly integrated into people's daily life. Followed by a large amount of video data on the Internet. It can be said that video data is an important medium with the most abundant information on the Internet, but also the most difficult to analyze. Since video is usually a kind of unstructured data without a fixed format, and usually has complex artistic expressions and exquisite shooting arrangements, how to automatically organize, store and retrieve a large amount of video data has naturally become the three major issues of video technology. a key question. [0003] Among various forms o...

AI Technical Summary

Problems solved by technology

However, film and television videos are very complex, and there are relatively few studies based on film and television videos.

[0007] 2. At present, many film and television research based on higher complexity are mostly methods related to the design of film and television content for small samples and small data volumes, but the performance of these methods on large-scale and complex data can hardly meet the actual needs (processing The speed is slow, and the requirements for training samples are more stringent)

It can be said that there are very few studies involving large-scale and complex film and television data, let alone the emergence of mature methods or frameworks

[0008] 3. Based on the extraction of key scenes, an important breakthrough direction is to "effectively bridge the semantic gap between the low-level features of the video and the high-level semantics of the video", but the semantic calculation of the video is not only slow but also inefficient. It seems that it is not very realistic to carry out commercial application, let alone meet the actual demand of rapid growth

Images