A video automatic labeling method and automatic labeling system based on motion matching

Abstract

The invention provides an automatic video labeling method based on motion matching. The method can automatically synchronize and label videos in real time, wherein the videos are shot in real time. The method is characterized by comprising the steps of collecting a target motion video of a motion target, collecting sensor data of an accessory motion sensor module when the motion target moves, extracting video motion features of the target motion video and sensor motion features of the sensor motion data, then carrying out feature matching analyzing on the video motion features and the sensor motion features, and finally labeling the target motion video according to the feature matching analysis result. An automatic video labeling method based on motion matching comprises the motion sensor module, a video collection module and a motion matching module, wherein the motion sensor module is used for collecting the sensor motion data of the motion target; the motion matching module and the motion sensor module are in data communication through a wireless communication module; the motion matching module and the video collection module are in data communication through a wireless communication module.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a video automatic labeling method and automatic labeling system based on motion matching. Background technique [0002] Video annotation is the description of video semantic content, which is the premise of video retrieval. In real life, it is often encountered that it is necessary to mark real-time shooting videos, such as real-time monitoring of specific targets while shooting specific targets in the video, and real-time shooting of tourist tours and commemorative videos in large playgrounds and scenic spots Annotation is carried out for tourists to retrieve and extract. However, most of the existing video annotations are based on training, matching and labeling of existing videos. At present, there is no automatic semantic description method for video content by using the target motion state. Contents of the invention [0003] In view of the above problems, the present in...

AI Technical Summary

Problems solved by technology

In real life, it is often encountered that it is necessary to mark real-time shooting videos, such as real-time monitoring of specific targets while shooting specific targets in the video, and real-time shooting of tourist tours and commemorative videos in large playgrounds and scenic spots Annotate for tourists to retrieve and extract. However, most of the existing video annotations are based on training, matching and labeling of existing videos. At present, there is no method for automatic semantic description of video content by using the target motion state.

Images