Residual neural network-based fully convolutional network for microstructure segmentation

被引：31

作者：

Jang, Junmyoung ^{[1
]}

Van, Donghyun ^{[1
]}

Jang, Hyojin ^{[1
]}

Baik, Dae Hyun ^{[2
]}

Yoo, Sang Duk ^{[2
]}

Park, Jaewoong ^{[1
]}

Mhin, Sungwook ^{[3
]}

Mazumder, Jyoti ^{[4
]}

Lee, Seung Hwan ^{[1
]}

机构：

[1] Korea Aerosp Univ, Sch Aerosp & Mech Engn, 76 Hanggongdaehang Ro, Goyang Si 10540, Gyeonggi Do, South Korea

[2] AIWARE, Seongnam Si, Gyeonggi Do, South Korea

[3] Korea Inst Ind Technol, Siheung Si, Gyeonggi Do, South Korea

[4] Univ Michigan, Ctr Laser & Plasmas Adv Mfg, Ann Arbor, MI 48109 USA

来源：

SCIENCE AND TECHNOLOGY OF WELDING AND JOINING | 2020年 / 25卷 / 04期

基金：

新加坡国家研究基金会;

关键词：

Submerged arc welding; carbon steel; acicular ferrite; fraction; segmentation; deep learning; fully convolutional network; ResNet; ACICULAR FERRITE; STEEL; MN;

D O I：

10.1080/13621718.2019.1687635

中图分类号：

T [工业技术];

学科分类号：

08 ;

摘要：

In this study, microstructures of weldment produced using carbon steel A516 grade 60 were analysed via a deep learning approach to measure the fraction of acicular ferrite which considerably influences on mechanical properties of carbon steel. The fully convolutional network was used to conduct the image segmentation. Submerged arc welding was used for welding, and the dataset was constructed using optical microscope. The model was compiled with ResNet, which is the state-of-the-art classifier used as an encoder. The model is trained to distinguish acicular ferrite from microstructures of dataset images and then estimate its accuracy. As a result, the mean intersection over union, which is a metric commonly used to evaluate image segmentation, was shown to be higher than 85%.

引用

页码：282 / 289

页数：8

共 33 条

[1]

ABSON D, 1988, IX153388 IIW

[2]

ABSON D, 1992, IX168792 IIW

[3]

Abson D., 1980, A scheme for quantitative description of ferritic weld metal microstructures

[4]

Adams B.L., Microstructure sensitive design for performance optimization