Real-time true and false motion judgment method based on long short-term memory network

Abstract

The invention discloses a real-time true and false motion judgment method based on a long short-term memory network, and the method comprises a model training stage: obtaining a data set: inputting a motion video into a human body key point detection model in a single-frame image form according to a sequence, outputting the key point data of a human body, and forming a data set sample; selecting a training set, inputting the training set into the LSTM + full-connection neural network, and finally calculating Loss and updating the Loss; the method further comprises an implementation judgment stage, using data to be detected as model input, outputting a judgment result, and the judgment result comprises the type of motion reflected in the data. The method has the beneficial effects that the model is established by utilizing the human body key point data based on the human body key point detection model, and the fitted model is used for identifying the human body motion type in the video and identifying whether the motion exists.

Description

technical field [0001] The invention relates to the technical field of data recognition, in particular to a method for judging real-time true and false motions based on a long-short-term memory network. Background technique [0002] As the country and society attach importance to the physical fitness of primary and middle school students, and the rapid development of artificial intelligence, it is inevitable that artificial intelligence will enter the field of sports. Whether the current motion is approximate for this motion calculation method. [0003] 1. Traditional image difference frame method [0004] The difference between the image passed in through the camera and the image passed in the previous frame is compared, and the difference is the moving part. [0005] Disadvantages: high cost, poor performance, high environmental requirements, and it is impossible to judge whether the exercise is really being done. [0006] 2. Deep learning classification (classification...

AI Technical Summary

Problems solved by technology

[0005] Disadvantages: high cost, poor performance, high environmental requirements, can not judge whether it is really doing the exercise

[0009] Disadvantages: high cost and poor performance

[0012] Disadvantages: high cost, poor performance, high environmental requirements, can not judge whether it is really doing the exercise

[0015] Disadvantages: Can't judge whether you are really doing the exercise

Images