Semi-supervised ladder network-based image hash retrieval method

A semi-supervised, image-based technology, applied in neural learning methods, biological neural network models, special data processing applications, etc., can solve problems such as low accuracy, long algorithm time-consuming, and insufficient sample labeling

Inactive Publication Date: 2018-07-31
NANJING UNIV OF POSTS & TELECOMM
View PDF5 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] In order to solve the technical problems raised by the above-mentioned background technology, the present invention aims to provide an image hash retrieval method based on a semi-supervised ladder network, which solves the problems of low accuracy, long time-consuming algorithm and insufficient sample labeling in the image retrieval process

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
  • Semi-supervised ladder network-based image hash retrieval method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0025] The technical solutions of the present invention will be described in detail below in conjunction with the accompanying drawings.

[0026] Such as figure 1 Shown, the implementation steps of the present invention are as follows.

[0027] Step 1, start, the detection system starts;

[0028] Step 2, determine the image data in the database;

[0029] Step 3, determining the image data to be retrieved;

[0030] Step 4, image data preprocessing, due to the use of deep learning to extract features, so the preprocessing is relatively simple, mainly including preprocessing operations such as mean removal and whitening;

[0031] Step 5, input the preprocessed image into a semi-supervised ladder network combined with a ladder network and a convolutional neural network (CNN) to obtain a feature map with a higher dimension;

[0032] In step 6, the feature map obtained in step 5 is expressed as a feature code containing only four values ​​of 0, 1, 2, and 3 by hashing. This featu...

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

The invention discloses a semi-supervised ladder network-based image hash retrieval method. The method comprises the steps of selecting images in a database and to-be-retrieved images to perform preprocessing; inputting the preprocessed images to a trained semi-supervised ladder network for performing feature extraction, wherein the semi-supervised ladder network comprises an unsupervised ladder network and a convolutional neural network, the convolutional neural network comprises a convolutional layer, a pooling layer, a full connection layer and a hidden layer connected in sequence, and thehidden layer performs hash expression on image features by utilizing a hash activation function to obtain feature codes; comparing the feature codes to determine a group which the to-be-retrieved images belong to; and calculating distances between the to-be-retrieved images and eigenvectors of the images in the group, and outputting N images with shortest distances. The problems of low accuracy, relatively long algorithm time and sample tagging deficiency in an image retrieval process are solved.

Description

technical field [0001] The invention belongs to the field of image retrieval, and in particular relates to an image hash retrieval method based on a semi-supervised ladder network. Background technique [0002] With the enhancement of hardware computing power, image retrieval has been more widely and deeply applied in various practical problems such as search engines, e-commerce, and video pedestrian re-detection. Due to the wide range of sources of images on the Internet, the amount of image data is extremely large, so it is impossible to give the best description for each image. For data processing, it is equivalent to more sample data, but the amount of labeled data is relatively small. . In this case, descriptions of similar images are beneficial for image understanding. Therefore, image search can solve this problem, that is, as long as there is an image with a high similarity to the image in the existing database, the image can be understood with the description of t...

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 Applications(China)
IPC IPC(8): G06F17/30G06K9/62G06N3/04G06N3/08
CPCG06F16/5838G06N3/088G06N3/045G06F18/22
Inventor 余瀚吴彬陈兴国
Owner NANJING UNIV OF POSTS & TELECOMM
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap