Dog Breed Classification Based on Deep Learning

Deep learning is part of the field of artificial intelligence. It has powerful feature extraction and learning capabilities. Because of its various advantages, it has been applied in many fields. Object detection is an important technology in deep learning, and object detection based on deep learning has also been studied by many people. With the gradual improvement of people's living standards, pets have gradually received people's love, among which pet dogs occupy the majority. Different types of pet dogs will bring different problems. For example, large pets may be more aggressive and cause problems for city management. If dangerous pets can be distinguished in time, it can bring more security to people and avoid some people being bitten by pet dogs. In the deep learning algorithm, YOLOv3 has better object detection performance, but it only targets different species and objects, and the classification of different categories of specific species is not good enough. In daily life, the body of the pet dog is sometimes hidden by the complicated background, which makes it difficult to extract the overall characteristics of the pet dog. At this time, the facial features of the pet dog can be fully utilized to distinguish the pet dog. In order to solve this problem, this paper proposes an improved yolov3 model for face detection and categorization of pet dogs. This paper establishes a data set of 8 different kinds of pet dogs. The data set is divided into training set and a test set, and the training set is sent to the established model for training. Finally, we use the test set to verify the effect of the model. This paper establishes a data set of 8 different kinds of pet dogs. Pet dog types include Akita, Golden Retriever, Poodle, Pomeranian, Samoyed, Corg, Pug, and Husky. Experiments show that this paper can realize the detection and classification of pet dogs with high detection speed and accuracy, and mAP(mean Average Precision) can reach 94.91%.