Image retrieval method based on depth feature consistent Hash algorithm

A hash algorithm and image retrieval technology, applied in the field of deep learning, can solve problems such as the image similarity ranking that cannot be well reflected, achieve retrieval accuracy and time advantages, improve retrieval performance, and improve the effect of loss function

Active Publication Date: 2021-08-31
OCEAN UNIV OF CHINA
View PDF11 Cites 5 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

For multi-label images, this method does not reflect the similarity ranking of images well

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
  • Image retrieval method based on depth feature consistent Hash algorithm
  • Image retrieval method based on depth feature consistent Hash algorithm
  • Image retrieval method based on depth feature consistent Hash algorithm

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0059] A kind of image retrieval method based on depth feature consistent hashing algorithm, comprises the following steps (such as figure 1 shown):

[0060] S1: First, get the semantic similarity matrix according to the label of the image data (such as figure 1 Semanticsimilarity matrix (semantic similarity matrix) part);

[0061] Given n training set images I={I 1 , I 2 , … , I n}, the value of n is a positive integer; first, the similarity matrix is ​​calculated using the labels. The traditional calculation method is, if I i and I j have any same label, then s ij =1, otherwise s ij =0. Follow the method of the predecessors, use the percentage to calculate s; the formula is as follows:

[0062] (1)

[0063] Among them, li and lj represent the label vectors of images Ii and Ij; represents the inner product of images Ii and Ij; according to formula (1), images are divided into two categories: strong similarity and weak similarity. Strong similarity is divided ...

Embodiment 2

[0095] In order to verify the effectiveness of the method, experiments were carried out on the widely used datasets Flickr and Cifar-10, and compared with other state-of-the-art methods. Flickr is a dataset containing 25,000 images, each image has at least a label. Resize the image to 227×227, an image may contain multiple labels. Cifar-10 is a color image dataset that is closer to general objects. Cifar-10 is a small dataset compiled by Hinton's students Alexkrizhevsky and Ilyasutskever to identify cosmic objects. There are 10 categories: Airplane, Car, Bird, Cat, Deer, Dog, Frog, Horse, Boat and Truck. The size of each image is 32×32, and each category has 6000 images. There are 50000 training images and 10000 testing images in the dataset.

[0096]For Flickr, 4000 images are randomly selected as the training set and 1000 images as the test set. Set λ=0.1, because λ will lead to more discretization, and too small value of λ will reduce the impact of the quantization los...

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 an image retrieval method based on a depth feature consistent Hash algorithm. The method comprises the following steps: acquiring multi-label or single-label image data including a training set and a test set; preprocessing the training set; optimizing the neural network by using the preprocessed training set; inputting the training set into the optimized neural network to obtain a hash code; and calculating Hamming distances between the Hash codes obtained by calculation and the Hash codes obtained by the test set, sorting according to the distances from small to large, and outputting the first k retrieval results to finish retrieval. Through verification, the model provided by the invention has better retrieval performance than other existing baseline methods. In the retrieval of single-label and multi-label image data sets, compared with the existing common method, the method disclosed by the invention has obvious advantages in retrieval precision and time.

Description

technical field [0001] The invention belongs to the technical field of deep learning, and in particular relates to a single-label and multi-label image retrieval method based on a deep feature consistent hashing algorithm. Background technique [0002] With the rapid development of multimedia big data, the number of images has exploded, which requires fast and accurate retrieval methods. Exact nearest neighbor retrieval (KNN) takes a long time and is not suitable for large data retrieval, while approximate nearest neighbor retrieval (ANN) is more popular due to the consideration of time and efficiency. [0003] Supervised learning is a common technique for training neural networks and decision trees. Both techniques, neural network and decision tree, are highly dependent on the information given by the predetermined classification system. For neural network, the classification system uses this information to judge the network error, and then continuously adjusts the network...

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): G06F16/51G06F16/583G06K9/62G06N3/04G06N3/08
CPCG06F16/51G06F16/583G06N3/08G06N3/048G06N3/045G06F18/22G06F18/241
Inventor 曹媛刘峻玮陶小旖桂杰
Owner OCEAN UNIV OF CHINA
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