Data retrieval method and system based on multi-image weighted fusion

Abstract

The invention belongs to the technical field of information retrieval and provides a data retrieval method and system based on multi-image weighted fusion. In order to solve the problem of inaccurate retrieval, the retrieval method includes obtaining the mapping matrix based on the objective function, and then projecting the test data according to the mapping matrix to generate the hash code matrix of the test data; calculating the hash code matrix of the test data and obtaining the hash code matrix based on the objective function. The Hamming distance between the hash code matrices of the training data is sorted to obtain the retrieval results of the test data; among them, the objective function is composed of six items, and the first two items are obtained by using the collaborative matrix decomposition to obtain different modal data Latent factor matrix; the third item is to learn the similarity graph matrix within and between modalities; the fourth item is to generate a unified hash code matrix through the unified consensus graph matrix and latent factor matrix between modalities; the fifth item is to learn the hash function; the sixth term is the regularization term. It has the characteristics of fast training and retrieval speed and high retrieval performance.

Description

technical field [0001] The invention belongs to the technical field of information retrieval, and in particular relates to a data retrieval method and system based on multi-image weighted fusion. Background technique [0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art. [0003] With the rapid development of network technology and the exponential growth of different modal data in social media, it is no longer limited to similarity retrieval between single-modal data, but more mutual retrieval between multi-modal data. The task of multimodal retrieval is to find a set of semantically similar objects in another modality given a query object in one modality, such as text retrieval image, image retrieval text, etc. However, in practical applications, data of different modalities often have inconsistent underlying feature structures, that is, the heterogeneity problem among mult...

AI Technical Summary

Problems solved by technology

However, in practical applications, data of different modalities often have inconsistent underlying feature structures, that is, the heterogeneity problem among multimodal data

At the same time, for large-scale databases, traditional multimedia retrieval technologies often cannot achieve accurate retrieval due to limitations in storage space and computing costs.

Images