Depth feature comparison weighted image retrieval method based on vector comparison strategy

Abstract

The invention discloses a depth feature comparison weighted image retrieval method based on a vector comparison strategy, and the method comprises the steps of inputting an image of a database into a convolutional neural network model, and extracting a depth feature; secondly, enhancing, fusing and comparing depth features, and forming and saving comparison weights; then, inputting a database image and a to-be-queried image into the convolutional neural network model, and extracting depth features; then, performing weighting, weighting enhancement and fusion on the depth features by using the comparison weight, and forming a matching feature vector; and finally, performing similarity matching on the matching feature vector of the to-be-queried image and the matching feature vector of the database image so as to return a retrieval image. According to the invention, a visual saliency comparison processing mechanism of a human brain is simulated, the depth features of the images are acquired by using the convolutional neural network model for comparison, distinguishable advanced semantic features of the images can be effectively described, and the image retrieval accuracy can be improved.

Description

technical field [0001] The invention relates to the technical field of image retrieval, in particular to a depth feature contrast weighted image retrieval method based on a vector comparison strategy. Background technique [0002] With the rapid development and wide application of smart phones and mobile communication technologies, people can now take pictures anytime and anywhere and upload them to the Internet through sharing channels very easily. In this context, the Internet has obtained a large amount of image data information. In the face of massive images, whether it is the enterprise manager on the server side or the individual on the user side, they all face the difficulties and challenges of retrieving images from the massive images. On the one hand, enterprise managers need to efficiently manage massive image data, among which image retrieval is one of their core management services, such as retrieving all images containing certain sensitive content in the image d...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the problem of how to simulate the mechanism of human brain's visual salience contrast processing information, the present invention provides a depth feature contrast weighted image retrieval method based on a vector comparison strategy

Images