Experimental mouse behavior analysis device and method based on YOLO and ST-GCN algorithm fusion

The invention discloses a behavior analysis device and method for laboratory mouse animal skeleton modeling based on YOLO and ST-GCN algorithms, when a user uses the device in a laboratory mouse animal feeding box and other places, a USB camera module capable of being connected with a mainboard is matched to collect images or videos of laboratory mouse animals and then sends the images or videos to a behavior analysis system. The system firstly realizes target detection and attitude estimation through a YOLO algorithm model, automatically captures dynamic skeleton points, accurately obtains skeleton point data of laboratory mouse animals, then carries out dynamic skeleton modeling on the data obtained by the previous model through an ST-GCN algorithm model, and finally carries out dynamic skeleton modeling on the data obtained by the previous model through the ST-GCN algorithm model. Behavior data such as behavior recognition, centroid data, limb length and motion trail of at least five motion modes such as t