Metallographic image segmentation of GCr15 bearing steel based on CGAN

Issue title: Selected papers from the International Symposium on Applied Electromagnetics and Mechanics - ISEM 2019

Guest editors: Jinhao Qiu, Ke Xiong and Hongli Ji

Article type: Research Article

Authors: Chen, Yuanyuan^{a; †} | Jin, Wuyin^a; | Wang, Meng^{b; †}

Affiliations: [a] School of Mechanical and Electronical Engineering, Lanzhou University of Technology, Lanzhou, Gansu, China | [b] School of Electronic and Information Engineering, Soochow University, Suzhou, Jiangsu, China

Correspondence: [*] Corresponding author: Wuyin Jin, School of Mechanical and Electronical Engineering, Lanzhou University of Technology, Lanzhou, Gansu, China. E-mail: [email protected]

Note: [†] Contributed equally.

Abstract: A novel deep learning segmentation method based on Conditional Generative Adversarial Nets (CGAN) is proposed, being U-GAN in this paper to overtake shortcomings of the metallographic images of GCr15 bearing steel, such as multi-noise, low contrast and difficult to segment. The results of experiment indicate that the proposed model is the most accurate comparing with the digital image processing methods and deep learning methods on carbide particle segmentation. The average Dice’s coefficient of similarity measure function is 0.9158, which is the state-of-the-art performance on dataset.

Keywords: Metallographic image, image processing, carbide particle segmentation, deep learning, CGAN

DOI: 10.3233/JAE-209441

Journal: International Journal of Applied Electromagnetics and Mechanics, vol. 64, no. 1-4, pp. 1237-1243, 2020