Label Noise Learning Method for Metallographic Image Recognition of Heat-Resistant Steel for Use in Pressure Equipment
In metallographic examination, spherular pearlite gradation, an important step in a metallographic examination, is the main indicator used to assess the reliability of heat-resistant steel. Recognition of pearlite spheroidization via the manual way mainly depends on the subjective perceptions and ex...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-10-01
|
| Series: | Materials |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1996-1944/15/19/7037 |
| _version_ | 1797478206734860288 |
|---|---|
| author | Zhiyuan Shen Haijun Hu Ziyi Huang Yu Zhang Yafei Wang Xiufeng Li |
| author_facet | Zhiyuan Shen Haijun Hu Ziyi Huang Yu Zhang Yafei Wang Xiufeng Li |
| author_sort | Zhiyuan Shen |
| collection | DOAJ |
| description | In metallographic examination, spherular pearlite gradation, an important step in a metallographic examination, is the main indicator used to assess the reliability of heat-resistant steel. Recognition of pearlite spheroidization via the manual way mainly depends on the subjective perceptions and experience of each inspector. Deep learning-based methods can eliminate the effects of the subjective factors that affect manual recognition. However, images with incorrect labels, known as noisy images, challenge successful application of image recognition of deep learning models to spherular pearlite gradation. A deep-learning-based label noise method for metallographic image recognition is thus proposed to solve this problem. We use a filtering process to pretreat the raw datasets and append a retraining process for deep learning models. The presented method was applied to image recognition for spherular pearlite gradation on a metallographic image dataset which contains 422 images. Meanwhile, three classic deep learning models were also used for image recognition, individually and coupled with the proposed method. Results showed that accuracy of image recognition by a deep learning model solely is lower than the one coupled with our method. Particularly, accuracy of ResNet18 was improved from 72.27% to 77.01%. |
| first_indexed | 2024-03-09T21:29:42Z |
| format | Article |
| id | doaj.art-95aefc20cb5b483792413b298e3d4dc1 |
| institution | Directory Open Access Journal |
| issn | 1996-1944 |
| language | English |
| last_indexed | 2024-03-09T21:29:42Z |
| publishDate | 2022-10-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Materials |
| spelling | doaj.art-95aefc20cb5b483792413b298e3d4dc12023-11-23T21:00:57ZengMDPI AGMaterials1996-19442022-10-011519703710.3390/ma15197037Label Noise Learning Method for Metallographic Image Recognition of Heat-Resistant Steel for Use in Pressure EquipmentZhiyuan Shen0Haijun Hu1Ziyi Huang2Yu Zhang3Yafei Wang4Xiufeng Li5School of Chemical Engineering, Xi’an Jiaotong University, Xi’an 710049, ChinaSchool of Chemical Engineering, Xi’an Jiaotong University, Xi’an 710049, ChinaSchool of Chemical Engineering, Xi’an Jiaotong University, Xi’an 710049, ChinaSchool of Computer Science, Shaanxi Normal University, Xi’an 710119, ChinaSchool of Chemical Engineering, Xi’an Jiaotong University, Xi’an 710049, ChinaChina Special Equipment Inspection and Research Institute, Beijing 100029, ChinaIn metallographic examination, spherular pearlite gradation, an important step in a metallographic examination, is the main indicator used to assess the reliability of heat-resistant steel. Recognition of pearlite spheroidization via the manual way mainly depends on the subjective perceptions and experience of each inspector. Deep learning-based methods can eliminate the effects of the subjective factors that affect manual recognition. However, images with incorrect labels, known as noisy images, challenge successful application of image recognition of deep learning models to spherular pearlite gradation. A deep-learning-based label noise method for metallographic image recognition is thus proposed to solve this problem. We use a filtering process to pretreat the raw datasets and append a retraining process for deep learning models. The presented method was applied to image recognition for spherular pearlite gradation on a metallographic image dataset which contains 422 images. Meanwhile, three classic deep learning models were also used for image recognition, individually and coupled with the proposed method. Results showed that accuracy of image recognition by a deep learning model solely is lower than the one coupled with our method. Particularly, accuracy of ResNet18 was improved from 72.27% to 77.01%.https://www.mdpi.com/1996-1944/15/19/7037heat-resistant steelmetallographic image recognitiondeep learninglabel noise learning |
| spellingShingle | Zhiyuan Shen Haijun Hu Ziyi Huang Yu Zhang Yafei Wang Xiufeng Li Label Noise Learning Method for Metallographic Image Recognition of Heat-Resistant Steel for Use in Pressure Equipment Materials heat-resistant steel metallographic image recognition deep learning label noise learning |
| title | Label Noise Learning Method for Metallographic Image Recognition of Heat-Resistant Steel for Use in Pressure Equipment |
| title_full | Label Noise Learning Method for Metallographic Image Recognition of Heat-Resistant Steel for Use in Pressure Equipment |
| title_fullStr | Label Noise Learning Method for Metallographic Image Recognition of Heat-Resistant Steel for Use in Pressure Equipment |
| title_full_unstemmed | Label Noise Learning Method for Metallographic Image Recognition of Heat-Resistant Steel for Use in Pressure Equipment |
| title_short | Label Noise Learning Method for Metallographic Image Recognition of Heat-Resistant Steel for Use in Pressure Equipment |
| title_sort | label noise learning method for metallographic image recognition of heat resistant steel for use in pressure equipment |
| topic | heat-resistant steel metallographic image recognition deep learning label noise learning |
| url | https://www.mdpi.com/1996-1944/15/19/7037 |
| work_keys_str_mv | AT zhiyuanshen labelnoiselearningmethodformetallographicimagerecognitionofheatresistantsteelforuseinpressureequipment AT haijunhu labelnoiselearningmethodformetallographicimagerecognitionofheatresistantsteelforuseinpressureequipment AT ziyihuang labelnoiselearningmethodformetallographicimagerecognitionofheatresistantsteelforuseinpressureequipment AT yuzhang labelnoiselearningmethodformetallographicimagerecognitionofheatresistantsteelforuseinpressureequipment AT yafeiwang labelnoiselearningmethodformetallographicimagerecognitionofheatresistantsteelforuseinpressureequipment AT xiufengli labelnoiselearningmethodformetallographicimagerecognitionofheatresistantsteelforuseinpressureequipment |