Unlock instant, AI-driven research and patent intelligence for your innovation.

Video interaction event analysis method and device based on sequence space-time cube features

A video interaction and cube technology, applied in the field of computer vision, can solve the problem of lack of description of local motion characteristics, and achieve the effect of intelligent detection and enhancement of description ability.

Inactive Publication Date: 2018-01-05
PEKING UNIV
View PDF4 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This analysis method lacks the description of the local motion characteristics of the object, and cannot determine the specific type of a complex event by analyzing the causal relationship between each stage of the complex event.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Video interaction event analysis method and device based on sequence space-time cube features
  • Video interaction event analysis method and device based on sequence space-time cube features
  • Video interaction event analysis method and device based on sequence space-time cube features

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0039] The present invention will be described in detail below through specific embodiments and accompanying drawings.

[0040] figure 1 It is a structural diagram of the video interaction event analysis device based on sequence space-time cube features of this embodiment, which includes: a preprocessing module for detecting and tracking objects of interest in surveillance videos; a video sequence division module connected to the preprocessing The processing module is used to adaptively divide the monitoring video into a space-time cube sequence based on the detection and tracking results; the space-time cube feature extraction module is connected to the preprocessing module and the video sequence division module, and is used to extract objects of interest in the monitoring video The visual features of the time-series feature reconstruction module, connected to the space-time cube feature extraction module, used to reconstruct the extracted space-time cube feature into a varia...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The present invention relates to a video interaction event analysis method and device based on sequence space-time cube features. The method includes: dividing the surveillance video into several space-time cube sequences based on the detection and tracking results of the surveillance video; extracting object trajectories in each space-time cube , appearance and local motion descriptors, and the extracted descriptors are composed of feature fragments; all feature fragments in the space-time cube are reconstructed to construct sequence space-time cube features, and the sequence space-time cube features are used to classify and detect interaction events. The device includes a preprocessing module, a video sequence division module, a space-time cube feature extraction module, a time series feature reconstruction module and a sequence feature classification module. The invention realizes the high-level semantic layer description of the monitoring video content, and uses the multi-core support vector machine based on the dynamic time alignment kernel function to realize the variable-length sequence feature classification, thereby realizing the intelligent detection of the interactive event of the monitoring video stream.

Description

technical field [0001] The invention belongs to the technical field of computer vision, and relates to a monitoring video interaction event analysis method, in particular to a monitoring video interaction event analysis method based on sequence space-time cube characteristics, and a device for realizing the method. Background technique [0002] As surveillance cameras are widely used in all aspects of people's lives, surveillance video data is growing explosively. How to intelligently analyze interesting events that occur in it is a challenging problem. Among the events of interest, some are multi-object interaction events, such as fights, robberies, homicides, and car crashes. In order to be able to report to the police in the early stage of the incident and assist in the investigation and evidence collection after the incident, the intelligent analysis method for such incidents is very critical. [0003] At present, the handling of such incidents mainly relies on witnesse...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/00G06T7/246
Inventor 田永鸿房晓宇王耀威黄铁军
Owner PEKING UNIV