Method, electronic device, and computer-readable storage medium for retrieving images

Abstract

According to embodiments of the present disclosure, a method for retrieving images, an electronic device, and a computer-readable storage medium are provided. The method includes: using a first neural network to process the image to be retrieved to determine a local feature vector of the image to be retrieved; using a second neural network to process the image to be retrieved to determine the global feature of the image to be retrieved vector; and based on the local feature vector and the global feature vector, determine an image whose similarity with the image to be retrieved is higher than a predetermined threshold.

Description

technical field [0001] The present disclosure relates to the field of image processing, and more particularly to methods, electronic devices, and computer-readable storage media for retrieving images. Background technique [0002] Deep learning is one of the most important breakthroughs in artificial intelligence in the last decade. It has achieved great success in speech recognition, natural language processing, computer vision, image and video analysis, multimedia and many other fields. For example, in traditional image retrieval techniques, the underlying visual features of images can usually be used for retrieval. However, due to the "semantic gap" problem between the underlying features and the high-level semantics, the image retrieval effect is not satisfactory. [0003] In contrast, in image retrieval technology based on deep learning, convolutional neural network (CNN) can be used for image retrieval, which has powerful learning ability and efficient feature expres...

AI Technical Summary

Problems solved by technology

However, due to the "semantic gap" problem between the underlying features and high-level semantics, the image retrieval effect is not satisfactory

Images