Teaching behavior analysis method and device based on target tracking and attitude detection

Abstract

The invention relates to a teaching behavior analysis method and device based on target tracking and attitude detection. The method comprises a step of acquiring a video stream and performing moving detection to obtain a window where a moving target is located, a step of detecting the target in the window and acquiring human hand and face targets, a step of establishing a multi-target tracker forthe acquired human hand and face targets and extracting a human body position according to the face target to carry out single-person gesture recognition, a step of judging a target behavior accordingto a tracking result of the multi-target tracker, performing behavior analysis according to a single-person gesture recognition result and carrying out outputting. According to the method and the device, the recognition of targets including the human face and hands based on a moving window is proposed. Since the speed of face recognition is much higher than that of usual gesture recognition, thegesture recognition is performed on the basis of the face recognition, and the efficiency and accuracy are greatly improved. The accuracy of the behavior analysis based on the target tracking is higher than that of the behavior analysis directly based on gesture recognition. In the single-person gesture recognition after the face recognition, the accuracy of gesture recognition is also greatly improved.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to a teaching behavior analysis method and device based on target tracking and attitude detection. Background technique [0002] Human behavior recognition and deep learning are research hotspots in the field of intelligent video analysis. In recent years, they have received extensive attention from academia and engineering circles. They are the theoretical basis for many fields such as intelligent video analysis and understanding, video surveillance, and human-computer interaction. In recent years, deep learning algorithms that have received widespread attention have been successfully applied to various fields such as speech recognition and graphic recognition. Deep learning theory has made outstanding achievements in the feature extraction of static images, and has been gradually extended to the research of video behavior recognition with time series. [0003] In the fi...

AI Technical Summary

Problems solved by technology

[0005] However, in the teaching scene, there are usually dozens of students appearing in the screen at the same time, and the lower body of the students is usually blocked by the desk;

Using the posture detection mentioned above to analyze the special behavior of students, the people in the whole screen will detect the posture, which will consume a lot of computing resources, fail to meet the real-time requirements, and easily cause misjudgment

Therefore, it cannot meet the behavior analysis in the teaching scene

Images