An End-to-End Image Retrieval System Based on Gravitational Field Deep Learning

In this paper, we design an end-to-end image retrieval system based on deep convolutional neural network (DCNN). Compared with the traditional method of using the deep convolutional activation features as the feature vector to match the image, we simplify the process of the algorithm and improve the problem of 'semantic gap' in the content-based image retrieval system. We first build an image matching database based on the gravitational field model, that is to add a similarity score label for each image in the database production phase. We then train the improved deep learning model and verify the effectiveness of the algorithm on the common image matching database (Caltech-101 and Holidays). Finally, the experimental results show that our improved deep learning model that used for image retrieval has excellent image matching ability. The overall retrieval accuracy inCaltech-101 and Holidays is 88.5% and 94.1%, respectively. As the number of returned images increases, the retrieval accuracy of the system decreases slightly and eventually becomes stable at a high value.