Video content analysis method and device for IPTV (Internet Protocol Television) supervision

Abstract

The invention provides a video content analysis method and device for IPTV (Internet Protocol Television) supervision. The method comprises the steps of: analyzing stability of a video content to be analyzed in time domain and space domain, and determining a target area, to be subjected to semantic recognition, in the video content; determining a feature point, which can represent the target area, in the target area according to the texture character of the target area, and computing the feature descriptor of the feature point; using the feature descriptor as a vision word to be annotated, performing matching treatment in a pre-generated visual tree search library, and thereby obtaining the semantic annotation of the feature descriptor, wherein the visual tree search library includes annotated vision words and semantic annotations of the same; determining the semantic annotation of the target area according to the semantic annotation of the feature descriptor. According to the video content analysis method and device, analysis of the video content with characteristics of diversity, complexity, real-time and the like can be achieved, and application requirements in the scene of IPTV supervision are met.

Description

【Technical field】 [0001] The present invention relates to the technical field of Internet Protocol Television (IPTV), in particular to a video content analysis method and device for IPTV supervision. 【Background technique】 [0002] As a new form of radio and television communication, IPTV provides users with radio and television services based on wide-area broadband networks through certain network protocols. Under this technical form, the quantity and size of video content are increasing exponentially, and at the same time, video content providers are diversified, which makes video content present a certain degree of diversity, complexity, and real-time performance. From the perspective of IPTV regulation, it is necessary to conduct an in-depth analysis of the ideology embodied in the regulated video content, and use the analysis results to help regulatory decisions. [0003] Existing video content analysis methods for IPTV monitoring scenarios are mainly scene detection t...

AI Technical Summary

Problems solved by technology

Scene detection technology uses the overall information in the scene to analyze the geographical information of the scene, which can provide the characteristics of the scene. Scene detection is a general analysis, and its analysis goal is not clear, and it cannot be given for the ideology embodied in the specific target in the video content. Specific analysis semantics, not suitable for IPTV supervision application scenarios

Images