Index and query method based on audio and video segment

Abstract

The invention relates to an index and query method based on audio-video segment, which belongs to the technology field of an audio-video application. The method comprises: scanning contents of audio-video documents and determining quantity, time and location of feature segment; description words and time of start and end of the feature segment; storing the index information of each feature segment in an index document; or directly storing the feature segment information in the tail or head of the audio-video contents; user inputs a word in query facilities or input by combining a plurality of words; returning all the corresponding feature segment information including queried words to user through description words; user finds needed audio-video contents according to the query results; precisely locating the needed feature segment. The method can realize precise and rapid query and improve query precision, which can precisely locate the start time of the feature segment according to the time information in the index and realize rapid browse of the feature segment.

Description

technical field [0001] The invention belongs to the technical field of audio and video applications, and in particular relates to an index and query method for audio and video clips, which can be used for services such as audio and video content search, audio and video advertisement placement, and audio and video content navigation. Background technique [0002] At present, audio and video content is widely used in various fields. With these applications, the amount of audio and video content has increased greatly. How can we quickly and efficiently query the required audio and video content? We can only rely on superior audio and video content indexing technology, so the audio and video content indexing technology is becoming more and more important. [0003] Most audio and video content is indexed according to the name of the audio and video content. When it is necessary to retrieve these audio and video content, directly query the name of the audio and video content. Thi...

AI Technical Summary

Problems solved by technology

[0006] There are already some formed products, but it is generally difficult to extract the features of video content. For example, when searching for audio and video content related to cars, there are many shapes of cars, and some cars are bionic designs, such as the Beetle. So how to judge whether a certain audio and video content is about the Beetle or about the Beetle car

At present, there are some methods to describe the video content, but the effect is not satisfactory.

[0007] The open audio and video search engines on the Internet include www.podscope.com and wWw.blinkx.com, but the search accuracy is not good, and the query efficiency is relatively low

Images