Image retrieval method and device

Abstract

The present invention provides an image retrieval method and device, which relate to the field of image retrieval. The method includes: receiving a query image sent by a user; obtaining confidence degrees corresponding to each category of the query image according to an image classification method; Retrieve the image features corresponding to each confidence category in the current confidence segment from the image feature library in order from large to small; gradually load the retrieved image features in the order of retrieval; sort the loaded image features ;Sending the processing results to the display interface step by step, so that the display interface can display, which avoids the situation of retrieving all image features in the established image database and then outputting the results, ensuring the rapidity of the retrieval result feedback time and stability; and there are fewer problems of missing real targets due to classification errors, which ensures the detection rate and recall rate of retrieval and other indicators, and is more suitable for application scenarios such as security video surveillance.

Description

technical field [0001] The present invention relates to the field of image retrieval, in particular to an image retrieval method and device. Background technique [0002] At present, the image retrieval methods in the prior art are all content-based massive image retrieval, which mainly includes two stages: 1) the image entry stage, which establishes a database of image features, specifically classifying the entry images according to their shapes; For each category, select representative sample pictures from the image library to form a sample library; extract the image features of all input images in the sample library and store them in the corresponding database; 2) query stage, which is image retrieval stage, specifically extracting the image features of the input query image, traversing the image features in the above database, and calculating the correlation between the traversed image features and the image features of the query image, and then displaying the images fro...

AI Technical Summary

Problems solved by technology

[0003] The inventor found in the research that the above-mentioned image retrieval method in the prior art is prone to misclassification and leads to the problem of missing the real target, and the above-mentioned method also needs to retrieve all the features in the database before outputting, so that the retrieval results Long feedback time and poor stability

Images