Behavior recognition method, device and equipment and readable storage medium

By using distortion correction and convolutional networks to extract features in video image processing, the problems of inaccurate behavior recognition and unreliable activity score in the prior art are solved, and high-accurate behavior recognition and activity score are achieved.

CN119723670APending Publication Date: 2025-03-28THE NAVAL MEDICAL UNIV OF PLA +1
0 Cites 0 Cited by

Patent Information

Application Number
CN202411905405.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-12-23
Publication Date
2025-03-28

AI Technical Summary

Technical Problem

The prior art lacks accurate methods in video image behavior recognition, cannot record the dynamic or static behavior of injured people for a long time and accurately, and lacks a reliable activity scoring system.

Method used

By acquiring video images, distortion correction is performed, image sequence is extracted, and spatial and temporal features are extracted from the spatial stream convolution network and the temporal stream convolution network respectively, feature fusion is performed, and behavior recognition results are obtained.

Benefits of technology

It improves the accuracy of behavior recognition and the quantitative accuracy of behavioral activity corresponding to video images, reduces the computational complexity of feature extraction, and provides a reliable activity scoring system.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention provides a behavior recognition method and device, equipment and a readable storage medium. The behavior recognition method comprises the following steps: acquiring a to-be-recognized video image; performing distortion correction processing on the video image to obtain a video image after distortion correction; extracting an image sequence corresponding to the video image; inputting the image sequence into a spatial stream convolutional network, and extracting spatial features corresponding to the image sequence in the spatial stream convolutional network; inputting the image sequence into a time flow convolutional network, and extracting time features corresponding to the image sequence in the time flow convolutional network; and carrying out feature fusion processing on the spatial features and the time features to obtain a behavior recognition result, in the behavior recognition method, based on different extraction modes, the time features and the spatial features of the video image are fully extracted, the calculation complexity of extracting the time features and the spatial features is reduced, the feature extraction accuracy is improved, and the behavior recognition efficiency is improved. And a more accurate behavior identification result can be obtained.
Need to check novelty before this filing date? Find Prior Art