Strong-correlation unsupervised cross-modal retrieval method guided by information amount

The invention relates to the technical field of cross-modal retrieval, in particular to a strong-correlation unsupervised cross-modal retrieval method guided by information amount, which is realized by the following steps of: firstly, extracting local features, global features and text features of a...

Full description

Saved in:
Bibliographic Details
Main Authors LI FANG, LUO XIAONAN, LAN RUSHI, DAI LIULIAN, YANG RUI
Format Patent
LanguageChinese
English
Published 15.09.2023
Subjects
CALCULATING
COMPUTING
COUNTING
ELECTRIC DIGITAL DATA PROCESSING
PHYSICS
Online AccessGet full text

Cover

More Information
Summary:The invention relates to the technical field of cross-modal retrieval, in particular to a strong-correlation unsupervised cross-modal retrieval method guided by information amount, which is realized by the following steps of: firstly, extracting local features, global features and text features of an image; enhancing local features and global features of the image; carrying out regularization processing on the enhanced local features; performing orthogonal fusion on the global features and the local features of the image by using an image feature fusion network; fusing the image features and the text features by using a multi-modal fusion network according to a different-modal feature information quantity conversion proportion principle; and finally, mapping different modal features into Hash codes, and carrying out similarity sorting by utilizing a Hamming distance so as to obtain a retrieval result. The method focuses on enhancement and fusion of data features, more semantic information can be obtained, and
Bibliography:Application Number: CN202310657100