Animal Image Retrieval Algorithm Based on Deep Learning

With the advent of big data and the continuous development of the Internet, traditional text-based image retrieval technology cannot meet people's needs, and has higher requirements for animal image detection. In the era of big data, data is a precious resource. Especially in the deep learning research, only a large number of image data sets can make the theoretical algorithm implementation possible. How to efficiently search from massive images and how to extract important features of images has become a hot research topic. This paper first briefly summarizes the research background of two image retrieval techniques and analyzes their advantages and disadvantages. It also introduces the development trend and research dynamics of image retrieval technology at home and abroad. Then it focuses on the knowledge and key technical aspects of content-based image retrieval and deep learning. Finally, the deep learning theory is integrated into the content-based image retrieval technology. It trains and constructs the deep convolutional network structure to image the animal images. Feature extraction, and then retrieved by the similarity algorithm. Compared with the classic content-based image retrieval technology, the image retrieval system designed in this paper eliminates the traditional image retrieval system to manually select image features. It can directly input sample data training, and the network can automatically learn and extract the image essence. feature. Moreover, the comparison of experimental results shows that the image retrieval algorithm based on deep convolution network has more efficient retrieval efficiency, with an average of more than 80%, which has higher retrieval efficiency than the classic image retrieval system. Key words: Deep Learning, Image Retrieval, Convolutional Neural Network, Animal Image