Mouse Behavior Discrimination Based on YOLOv7-Pose

In animal stimulation experiments, accurately identifying and quantifying directional movement patterns such as left and right turning behaviors in mice is essential for evaluating neurological responses and stimulus effectiveness. Traditional manual observation methods are time-consuming and prone to subjective bias, while existing automated systems often lack precision in detecting subtle directional changes. This study optimized the YOLOv7-pose algorithm, improving accuracy by 3%, for detecting key points in mice. By analyzing the relative positions of five key points, specific behaviors were identified, and turning amplitude was quantified using the angle of the middle body segment. Finally, a user-friendly visualization interface was designed using QT.