Sample data enhancement method and system based on GAN network

A technology of sample data and sample data sets, applied in neural learning methods, biological neural network models, instruments, etc., can solve problems such as economic losses, performance degradation, and material environmental pollution

Pending Publication Date: 2022-01-21
XI AN JIAOTONG UNIV +1
View PDF0 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0002] The aging of polymer materials has become a very important problem, and the actual harm caused is much more serious than people imagined, especially under harsh environmental conditions, often leading to premature failure of equipment and a large loss of materials, not only economically affected Great losses, resulting in waste of resources, and even environmental pollution due to failure and decomposition of materials
However, during the process of processing, storage and use, polymer materials will be degraded under the comprehensive action of internal and external factors such as light, heat, water, chemical and biological erosion, and the performance will gradually decline, thus partially losing or losing their use value.

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
  • Sample data enhancement method and system based on GAN network
  • Sample data enhancement method and system based on GAN network
  • Sample data enhancement method and system based on GAN network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0067] The sample data enhancement method based on the GAN network in this embodiment includes the following steps:

[0068] Step 1: Collect a small sample dataset of aging samples

[0069] The insulation sample is peeled off from the stator bar to obtain single-layer or multi-layer stacked insulation sheets. Cut the insulating sheet into a sample of 1 cm×1 cm. The samples are aged to obtain a dataset of 117 aged samples as input images.

[0070] Step 2, constructing a pyramid-shaped GAN network learning model composed of N GAN network structures;

[0071] The model includes a pyramid-structured GAN, and training and inference are performed from coarse to fine. At each scale, G n learns to generate image samples, while the discriminator D n Unable to distinguish all resampled image patches from downsampled training image patches x n ; As the pyramid goes up, the effective patch size decreases (yellow area in the original image). G n The input of is a random noise image...

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 sample data enhancement method and system based on a GAN network. The method specifically comprises the following steps: constructing a pyramid-shaped GAN network learning model composed of N GAN network structures; training and testing a GAN network learning model based on a Coarse-to-Fine thought, and carrying out multiple iterations on an input image from a rough resolution; when an iteration result converges, adding an additional convolutional layer to increase the size of a generator, and adding residual connection from the original up-sampling feature to the output of the newly added convolutional layer until the resolution of the image reaches a set output resolution; and generating a virtual sample based on the image meeting the output resolution, and mixing the virtual sample with the small sample data set to obtain enhanced complete sample data. The problem that aging sample data are scarce in the industrial process is solved. Compared with a traditional data enhancement method, the method of the invention is advantageous in that: the efficiency is higher, and the data cost is reduced; and the generated image is more diversified on the basis of conforming to the original distribution.

Description

technical field [0001] The invention belongs to the technical field of GAN network models, in particular to a sample data enhancement method and system based on a GAN network. Background technique [0002] The aging of polymer materials has become a very important problem, and the actual harm caused is much more serious than people imagined, especially under harsh environmental conditions, often leading to premature failure of equipment and a large loss of materials, not only economically affected A great loss leads to the waste of resources, and even the pollution of the environment due to the failure and decomposition of materials. However, during the process of processing, storage and use, polymer materials will be degraded under the comprehensive action of internal and external factors such as light, heat, water, chemical and biological erosion, and the performance will gradually decline, thus partially losing or losing their use value. . [0003] Research on aging sam...

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): G06N3/04G06N3/08G06K9/62G06V10/774G06V10/82
CPCG06N3/08G06N3/045G06F18/214
Inventor张跃刘伟胡波梁智明唐丽汪建基
OwnerXI AN JIAOTONG UNIV