video behavior identification method based on a neural network and terminal equipment

The method is applicable to the technical field of neural networks. The invention provides a video behavior identification method based on a neural network and terminal equipment. The method comprisesthe following steps: obtaining a sample; cutting the video data into first video segments, continuously extracting a first picture with a first preset frame number from each first video segment, extracting a behavior feature part from each first picture to obtain a second picture; converting the extracted second picture into a second video segment; wherein the behavior feature part is an image ofan object with a changed action in a first picture; extracting a third picture from each second video segment according to a preset step length, converting the third picture into a third video segment, and finally importing each third video segment into a preset neural network model to obtain a behavior classification coefficient of each third video segment; According to the behavior classification coefficient of each thi