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

Lightweight video behavior identification method based on long-short-term time domain modeling algorithm

A long-term and short-term, modular algorithm technology, applied in the field of video classification and computer vision, can solve the problems of inaccurate recognition results and high consumption of computing resources, and achieve the effect of small amount of calculation and high accuracy of behavior recognition

Active Publication Date: 2020-07-10
NORTHWESTERN POLYTECHNICAL UNIV
View PDF14 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Compared with the two-stream algorithm and the three-dimensional convolutional neural network algorithm, the present invention effectively solves the problems of inaccurate recognition results and high consumption of computing resources in the current video behavior recognition technology without the need for additional training graph models and extraction of optical flow features

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
  • Lightweight video behavior identification method based on long-short-term time domain modeling algorithm
  • Lightweight video behavior identification method based on long-short-term time domain modeling algorithm
  • Lightweight video behavior identification method based on long-short-term time domain modeling algorithm

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0016] The present invention will be further described below in conjunction with the accompanying drawings and embodiments, and the present invention includes but not limited to the following embodiments.

[0017] The present invention provides a lightweight video behavior recognition method based on a long-term and short-term time-domain modeling algorithm, and its implementation steps are as follows:

[0018] 1. Video preprocessing

[0019] For the videos in the data set, first extract 8 frames of video clips from each video by uniform sampling method, and then perform multi-scale cropping (such as center cropping, etc.) on the video clips, and convert the size of each frame to 224× 224, so that each video is converted into a video segment with a size of 8×3×224×224. All the video clips constitute a new video clip dataset, and the tags of the videos in the original dataset are used as the tags of the corresponding video clips in the new video clip dataset. Finally, the new...

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 invention provides a lightweight video behavior identification method based on a long-short-term time domain modeling algorithm. A short-term feature interchange module is constructed by using a partial channel interchange method, a long-term feature fusion module is constructed by using graph convolution, effective extraction of short-term and long-term time features of the video is realizedrespectively, and the time features of different stages are extracted by inserting the two modules into different positions of the two-dimensional deep residual network, so that the problems of inaccurate result and high computing resource consumption of a current video behavior recognition technology are effectively solved.

Description

technical field [0001] The invention belongs to the technical field of computer vision and video classification, and specifically relates to a lightweight video behavior recognition method based on a long-term and short-term time-domain modeling algorithm, which can be applied to intelligent monitoring, crowd analysis, human-computer interaction and the like. Background technique [0002] With the emergence of short video software such as Douyin and Kuaishou and some live broadcast platforms, a large number of new videos are generated and shared on the Internet almost every moment. To cope with this information explosion, it becomes increasingly important to analyze and understand video information applied to various scenarios. Video behavior recognition refers to the recognition and judgment of human behavior in video, which has a wide range of applications in real life. is still a very challenging task. [0003] Video behavior recognition technology can classify the curr...

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
IPC IPC(8): G06K9/00G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06V20/46G06N3/045G06F18/253
Inventor 王琦李学龙白思开
Owner NORTHWESTERN POLYTECHNICAL UNIV