A method and an apparatus for general image classification base on a semi-supervised generative adversarial network

A semi-supervised, network technology, applied in the field of deep learning, can solve problems such as large deformation, and achieve the effect of improving accuracy, generalization performance and classification accuracy

Pending Publication Date: 2019-01-11
SHANGHAI MUNICIPAL ELECTRIC POWER CO +1
View PDF3 Cites 15 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The image size of MNIST is 28*28, that is, 784 dimensions. The hand

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 method and an apparatus for general image classification base on a semi-supervised generative adversarial network
  • A method and an apparatus for general image classification base on a semi-supervised generative adversarial network
  • A method and an apparatus for general image classification base on a semi-supervised generative adversarial network

Examples

Experimental program
Comparison scheme
Effect test

example 1

[0079] Example 1 General image classification method based on semi-supervised generative confrontation network

[0080] Taking the MNIST handwritten digit data set as an example, according to the method of the present invention, compare the probability under the semi-supervised method and the supervised method, and the classification accuracy obtained is as shown in Table 1:

[0081]Table 1. Classification accuracy using semi-supervised learning and supervised learning on the MNIST dataset

[0082] Number of labeled images

[0083] In Table 1, the last row represents the classification accuracy obtained by the supervised learning method, and the remaining rows represent the classification accuracy obtained by the semi-supervised learning method of the present invention, and "the number of labeled images" refers to the original training data. 50,000 is the size of the MNIST training data set. In the experiment, keep the input size of supervised learning and semi-super...

example 2

[0084] Example 2 General image classification method based on semi-supervised generative confrontation network

[0085] Taking the CIFAR-10 data set as an example, according to the method of the present invention, comparing the probability under the semi-supervised method and the supervised method, the obtained classification accuracy is shown in Table 2:

[0086] Table 2. Classification accuracy using semi-supervised learning and supervised learning on the MNIST dataset

[0087]

[0088]

[0089] In Table 2, the last row represents the classification accuracy rate obtained by the supervised learning method, and the remaining rows represent the classification accuracy rate obtained by the semi-supervised learning method of the present invention. "The number of labeled images" refers to the original training data. 50,000 is the size of the CIFAR-10 training data set. In the experiment, keep the input size of supervised learning and semi-supervised learning equal to get th...

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 method and an apparatus for general image classification base on a semi-supervised generative adversarial network are disclosed, relating to the image classification technology. The method includes:1, obtaining a deep convolutional generative adversarial network DCGAN through training, wherein that DCGAN comprises a generation network and a discrimination network, and the discrimination networkcomprises a convolutional neural network and a Softmax multi-classifier connected in turn; 2, inputting the image to be classified into the convolution neural network to obtain image features; 3, inputting the obtained image features into the Softmax multi-classifier to obtain a classification result. As the discriminant network is train in the DCGAN, the generalization performance and the classification accuracy rate of the discriminant network are improved. After image features are obtained through the discriminant network and Softmax multi-classifier is combined, the accuracy rate of imageclassification can be greatly improved.

Description

technical field [0001] The invention belongs to the field of deep learning and relates to image classification technology, in particular to a general image classification method and device based on a semi-supervised generation confrontation network. Background technique [0002] Image classification technology is one of the main branches in the field of computer vision and pattern recognition. Image classification is an image processing method that distinguishes different types of objects in the image according to the different characteristics reflected in the image information. Image classification is the use of computers to conduct quantitative analysis of images, and to classify an image or a certain area in an image into one of several categories to replace human visual interpretation. With the advent of the big data era, data is becoming more and more obvious in computer vision tasks. When there is enough data, basic models and algorithms can be used, such as KNN (k-Ne...

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
IPC IPC(8): G06K9/62
CPCG06F18/2413
Inventor 苏磊凌平张万才
Owner SHANGHAI MUNICIPAL ELECTRIC POWER CO
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