A Method for Segmenting Paint Cracks on ICGA Images Based on Conditional Generative Adversarial Networks

Abstract

The invention discloses a method for segmenting paint cracks on an ICGA image based on a conditional generative confrontation network. After normalization processing between the contrast images and the gold standard, they are spliced ​​into a set of images as sample data, and the samples are distributed in proportion as training set and test set; (2) Based on the principle of conditional generative confrontation network, construct generator and discriminant (3) Input the training set data into the network for confrontation training, define the loss function, and train the generator to generate the paint crack image corresponding to the original image; (4) In the test phase, input the test set data, and pass the trained generator G, get the corresponding paint crack segmentation result map. The segmentation method provided by the present invention can be used to solve the problems of less ICGA image samples and difficulty in obtaining contrast images, and has the characteristics of high accuracy of segmentation results.

Description

technical field [0001] The invention relates to a method for segmenting paint cracks on an ICGA image based on a conditional generative confrontation network, and belongs to the technical field of image processing. Background technique [0002] In recent years, with the rapid development of big data, deep learning networks have been widely used in computer vision, artificial intelligence and other fields. Among them, the generative confrontation network (GAN) is an important network tool to solve the problem of image translation. It is called "the coolest idea in the field of machine learning in the past 20 years". [0003] Compared with traditional graphical models, GAN is a better generative model, which in a sense avoids the Markov chain learning mechanism, which makes it different from traditional probabilistic generative models. Traditional probability generation models generally require Markov chain sampling and inference, but GAN avoids this process of extremely high...

AI Technical Summary

Problems solved by technology

[0005] The most important point is that when the probability density cannot be calculated, some generative models that traditionally rely on the natural interpretation of data cannot be learned and applied on it.

Images