Fast multi-label picture retrieval system and realization method

Abstract

The invention discloses a fast multi-label picture retrieval system and a realization method. The method comprises the following steps: deploying an RPN (Region Proposal Network) for extracting region proposals in a convolutional neural network, extracting region proposal information of pictures, and performing ROI pooling calculation on the region proposal information; after pooling, building a multi-label classification loss function through a fully connected layer according to multi-label information to train the convolutional neural network, and building a weighted three-dimensional loss function to train the convolutional neural network; extracting the hash code of each picture from a picture candidate set through the convolutional neural network after multi-task learning, saving the hash codes to a database, and comparing the hash codes with the hash codes in the database, thus completing picture retrieval. The whole network is trained through multi-task learning of classification and hashing, and therefore, the accuracy of retrieval is ensured. Moreover, the similarity is measured using Hamming distance in the process of retrieval, and the efficiency of retrieval is improved greatly.

Description

technical field [0001] The invention relates to the fields of neural convolution network and image processing, in particular to a fast multi-label image retrieval system and its implementation method. Background technique [0002] Nowadays, for multi-label image retrieval systems, in order to improve the speed of retrieval, many methods use the hash method to binarize the features, and use the Hamming distance to measure the similarity. In terms of hashing methods, there are mainly two types: unsupervised and supervised. [0003] Unsupervised methods, such as LSH, form hash functions through random mapping or random permutation, and do not depend on data points in space. This method often requires a longer hash code to obtain better performance. [0004] The supervised method is to learn the hash function from the training set through the learning method, and utilizes the labeling information in the training set. However, most of the current methods are single-task learning...

AI Technical Summary

Problems solved by technology

[0004] The supervised method is to learn the hash function from the training set through the learning method, using the labeling information in the training set, but most of the current methods are single-task learning, that is, only using paired labeling information

Images