Image data multi-label classification method

A classification method and image data technology, applied in the field of machine learning, can solve problems such as classification performance degradation

Pending Publication Date: 2020-10-20
ZHEJIANG UNIV
View PDF0 Cites 5 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

But when an image instance has multiple related labels, simple label propagation cannot make good use of this correlation, resulting in a decline in classification performance.

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 data multi-label classification method
  • Image data multi-label classification method
  • Image data multi-label classification method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0057] The collection of pictures to be classified D={pic i |1≤i≤n}, where n is the total number of pictures in the picture set D. In the picture set D, there are q ordered labels to be marked on each picture, and a total of l pictures have been marked whether they have the q labels, and the marked picture is pic i , 1≤i≤l.

[0058] (1) Calculate the weight matrix W and construct the graph relationship between pictures. This step includes the following sub-steps:

[0059] (1.1) Use the public dataset ImageNet on the website http: / / www.image-net.org / to train according to the method provided in 3.1 of the paper VERY DEEP CONVOLUTIONAL NETWORKS FOR LARGE-SCALE IMAGE RECOGNITION jointly published by Karen Simonyan and Andrew Zisserman in 2015 Change the convolutional neural network VGG Net proposed in this paper. VGG Net was jointly developed by the Visual Geometry Group of Oxford University and researchers from Google DeepMind, which can extract the feature vectors of the pic...

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 data multi-label classification method. The method comprises the following steps: firstly, constructing a topological relation of pictures; then, using a multi-label propagation algorithm (CMLP) based on cooperation for solving the problem of semi-supervised multi-label learning (SSML), thereby obtaining a confidence matrix of unmarked pictures. According to the invention, the multi-label classification prediction performance of the pictures can be improved by fully utilizing the correlation among the plurality of labels of the multi-label image data, so that the accuracy and robustness are improved; the distribution information of the non-supervised image data can be mined, so that the method well solves the problem of semi-supervised multi-label learning,and is superior to the existing semi-supervised image data multi-label classification method in performance.

Description

technical field [0001] The invention relates to the field of machine learning, in particular to a multi-label classification method for image data. Background technique [0002] Image classification is one of the most widely used machine learning problems. In real-world image classification applications, an image instance is often associated with multiple labels, and there are correlations among these labels. Therefore, classifying images with multiple related labels is an important research task. In traditional image labeling problems, a common assumption is that each image in the training dataset is accurately labeled. Unfortunately, in many real problems, although a large number of images are easy to obtain, the labels of the images need to go through an expensive and time-consuming manual labeling process. [0003] To this end, people have proposed a semi-supervised learning method, adding a large number of unlabeled images to limited labeled images to train together ...

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): G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06N3/045G06F18/24147G06F18/2411G06F18/214
Inventor 陈刚胡天磊陈珂刘雨辰李梦谨王皓波寿黎但
Owner ZHEJIANG UNIV
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