{"title":"通过平滑单热编码改进铸件数字射线照片中的散射缺陷分级","authors":"Han Yu, Xingjie Li, Xue Hao, Zhaowei Song, Shangyu Liu, Xinyue Li, Chunyu Hou, Huasheng Xie","doi":"10.1007/s40962-024-01335-3","DOIUrl":null,"url":null,"abstract":"<p>Ensuring the precise grading of discontinuities is imperative to guarantee the quality of castings and enhance profitability in casting production. Recent grading methods leveraging computer vision are advanced by performing a single-label image classification or regression, which loses the intrinsically ordinal relationship. Motivated by this observation, we propose a label smoothing technology for ordinal variables to convert the level of each defect instance into a discrete probability distribution, aiming to model the noise label and ordinal relationship. Furthermore, we design a convolutional neural network framework based on multi-task learning. This framework, by simultaneously learning the level label distribution and regressing the level directly, outperforms a single-task network in terms of overall performance. Finally, we construct a casting gas porosity defect grading dataset. Experimental results on this dataset highlight the significant advantages of our proposed method compared to traditional single-label image classification or regression algorithms.</p>","PeriodicalId":14231,"journal":{"name":"International Journal of Metalcasting","volume":"28 1","pages":""},"PeriodicalIF":2.6000,"publicationDate":"2024-04-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Improving Scattered Defect Grading in Castings Digital Radiographs via Smoothing the One-Hot Encoding\",\"authors\":\"Han Yu, Xingjie Li, Xue Hao, Zhaowei Song, Shangyu Liu, Xinyue Li, Chunyu Hou, Huasheng Xie\",\"doi\":\"10.1007/s40962-024-01335-3\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p>Ensuring the precise grading of discontinuities is imperative to guarantee the quality of castings and enhance profitability in casting production. Recent grading methods leveraging computer vision are advanced by performing a single-label image classification or regression, which loses the intrinsically ordinal relationship. Motivated by this observation, we propose a label smoothing technology for ordinal variables to convert the level of each defect instance into a discrete probability distribution, aiming to model the noise label and ordinal relationship. Furthermore, we design a convolutional neural network framework based on multi-task learning. This framework, by simultaneously learning the level label distribution and regressing the level directly, outperforms a single-task network in terms of overall performance. Finally, we construct a casting gas porosity defect grading dataset. Experimental results on this dataset highlight the significant advantages of our proposed method compared to traditional single-label image classification or regression algorithms.</p>\",\"PeriodicalId\":14231,\"journal\":{\"name\":\"International Journal of Metalcasting\",\"volume\":\"28 1\",\"pages\":\"\"},\"PeriodicalIF\":2.6000,\"publicationDate\":\"2024-04-22\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"International Journal of Metalcasting\",\"FirstCategoryId\":\"88\",\"ListUrlMain\":\"https://doi.org/10.1007/s40962-024-01335-3\",\"RegionNum\":3,\"RegionCategory\":\"材料科学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q2\",\"JCRName\":\"METALLURGY & METALLURGICAL ENGINEERING\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"International Journal of Metalcasting","FirstCategoryId":"88","ListUrlMain":"https://doi.org/10.1007/s40962-024-01335-3","RegionNum":3,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"METALLURGY & METALLURGICAL ENGINEERING","Score":null,"Total":0}
Improving Scattered Defect Grading in Castings Digital Radiographs via Smoothing the One-Hot Encoding
Ensuring the precise grading of discontinuities is imperative to guarantee the quality of castings and enhance profitability in casting production. Recent grading methods leveraging computer vision are advanced by performing a single-label image classification or regression, which loses the intrinsically ordinal relationship. Motivated by this observation, we propose a label smoothing technology for ordinal variables to convert the level of each defect instance into a discrete probability distribution, aiming to model the noise label and ordinal relationship. Furthermore, we design a convolutional neural network framework based on multi-task learning. This framework, by simultaneously learning the level label distribution and regressing the level directly, outperforms a single-task network in terms of overall performance. Finally, we construct a casting gas porosity defect grading dataset. Experimental results on this dataset highlight the significant advantages of our proposed method compared to traditional single-label image classification or regression algorithms.
期刊介绍:
The International Journal of Metalcasting is dedicated to leading the transfer of research and technology for the global metalcasting industry. The quarterly publication keeps the latest developments in metalcasting research and technology in front of the scientific leaders in our global industry throughout the year. All papers published in the the journal are approved after a rigorous peer review process. The editorial peer review board represents three international metalcasting groups: academia (metalcasting professors), science and research (personnel from national labs, research and scientific institutions), and industry (leading technical personnel from metalcasting facilities).