Unlock instant, AI-driven research and patent intelligence for your innovation.

A cross-modal hash retrieval method based on deep learning

A deep learning, cross-modal technology, applied in the direction of still image data retrieval, unstructured text data retrieval, text database indexing, etc., can solve the problem of not being able to mine the original feature identification information well, and achieve the promotion of mining and Retrieve performance, improve performance, promote the effect of improvement

Active Publication Date: 2022-06-03
JIUJIANG UNIV
View PDF10 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] The purpose of the present invention is to provide a cross-modal hash retrieval method based on deep learning, which solves the problem that the existing cross-modal hash retrieval method based on shallow learning structure cannot well mine the identification of original features. information problem

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A cross-modal hash retrieval method based on deep learning
  • A cross-modal hash retrieval method based on deep learning
  • A cross-modal hash retrieval method based on deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

Meaning:

[0048]

where, for each modality, in order to make the local neighbor structure of the data point in the Hamming space

It is consistent with the original feature space, that is: make each data point in the original feature space and its neighbor relationship in the Hamming space

is maintained, the following objective function can be designed:

[0050]

Based on the class label information of the object, the data points v of the image modality can be defined

i

(i=1,2,...,n) and text

modal data point t

j

(j=1,2,...,n) the semantic affinity matrix shown below:

[0052]

It should be noted that: as long as v

i

and t

j

belong to at least one of the same categories, they are considered to have the same language

righteous. To maintain inter-modal consistency between image modalities and text modalities in Hamming space, the following objectives can be designed

function:

[0054]

[0055] Based on the above, about image modal depth feature learning,...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

A cross-modal hash retrieval method based on deep learning, assuming that the pixel feature vector set of the image modality of an object is , the feature is that the method includes the following steps: (1) using the objective function designed based on deep learning technology to obtain Binary hash coding shared by image modality and text modality, sum of deep neural network parameters of image modality and text modality, and projection matrix sum of image modality and text modality; (2) Solve using alternate update method The unknown variables in the objective function , , , and; (3) The sum of the deep neural network parameters based on the solved image modality and text modality, and the projection matrix sum; (4) Calculate the query sample based on the generated binary hash code Hamming distance to each sample in the retrieval sample set; (5) Use a cross-modal retriever based on approximate nearest neighbor search to complete the retrieval of query samples. This method effectively improves the performance of cross-modal hash retrieval.

Description

A cross-modal hash retrieval method based on deep learning technical field [0001] The present invention relates to a cross-modal hash retrieval method based on deep learning. Background technique [0002] With the rapid development of science and technology and social productivity, the era of big data has come quietly. Big data is A collection of data that cannot be captured, managed, and processed using conventional software tools within a certain time frame. IBM offers Big data has 5V characteristics, namely: Volume (large amount of data), Variety (variety of types and sources), Value (data price) Value density is relatively low and sometimes precious), Velocity (data grows fast), Veracity (quality of data) quantity). Big data can also be considered as the need for new processing models to have stronger decision-making power, insight discovery and process optimization capability information assets. Information retrieval is an important aspect of data processin...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06F16/31G06F16/51
CPCG06F16/325G06F16/51
Inventor 董西伟邓安远周军杨茂保孙丽胡芳贾海英王海霞
Owner JIUJIANG UNIV