A Pet Detection System Based on YOLOv4

With the increasing development of artificial intelligence, it brings an opportunity to use advanced intelligent technology to solve real pet problems. And it has important practical significance to solve a series of issues such as pet photography and public transportation pet detection in a faster and more efficient way. By studying the application of deep convolutional neural networks in pet detection tasks, a complete system for pet detection is designed. The entire system uses the YOLOv4 algorithm as the basic algorithm for object detection. After completing the process of data collection, data expansion and data labeling, completing the algorithm training and optimization process, quantitatively analyzing the final system detection effect and testing the robustness and generalization of the system, a system for cat and dog detection with a mean average precision of 95.71% is finally obtained. Experiments show that the designed detection system can use the deep convolutional neural network to automatically, quickly and accurately detect pets.