{"title":"Res2-UNet++:用于电阻断层扫描的深度学习图像后处理方法","authors":"Qiushi Huang, Guanghui Liang, Chao Tan, Feng Dong","doi":"10.1088/1361-6501/ad57e0","DOIUrl":null,"url":null,"abstract":"\n It is challenging to monitor the multiphase flow distribution in the industrial processes in order to optimize the production. Electrical resistance tomography (ERT) can be used to visualize the inner distribution of multiphase flow. The image reconstruction plays a vital role in ERT. However, the nonlinearity and ill-posedness of inverse problem make the image reconstruction of ERT a challenge, and the development of advanced imaging algorithm has attracted much attention in the past. In this work, an improved U-shaped deep learning model is proposed, which combines the advantages of multi-scale feature extraction of UNet++ and residual feature fusion of Res2Net. The network is used to post-process the prereconstruction result of traditional ERT image reconstruction methods, where the generalization ability of the traditional methods and the flexible feature extraction advantage of deep learning methods can be combined. Simulations and experiments are designed to verify the generalization ability and the effectiveness of the proposed model. Both simulation and experimental results show that the proposed U-shaped network approach outperforms other deep learning methods, and the proposed deep learning model is fit for post-processing tasks of ERT.","PeriodicalId":2,"journal":{"name":"ACS Applied Bio Materials","volume":"20 4","pages":""},"PeriodicalIF":4.6000,"publicationDate":"2024-06-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Res2-UNet++: A Deep Learning Image Post-Processing Method for Electrical Resistance Tomography\",\"authors\":\"Qiushi Huang, Guanghui Liang, Chao Tan, Feng Dong\",\"doi\":\"10.1088/1361-6501/ad57e0\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"\\n It is challenging to monitor the multiphase flow distribution in the industrial processes in order to optimize the production. Electrical resistance tomography (ERT) can be used to visualize the inner distribution of multiphase flow. The image reconstruction plays a vital role in ERT. However, the nonlinearity and ill-posedness of inverse problem make the image reconstruction of ERT a challenge, and the development of advanced imaging algorithm has attracted much attention in the past. In this work, an improved U-shaped deep learning model is proposed, which combines the advantages of multi-scale feature extraction of UNet++ and residual feature fusion of Res2Net. The network is used to post-process the prereconstruction result of traditional ERT image reconstruction methods, where the generalization ability of the traditional methods and the flexible feature extraction advantage of deep learning methods can be combined. Simulations and experiments are designed to verify the generalization ability and the effectiveness of the proposed model. Both simulation and experimental results show that the proposed U-shaped network approach outperforms other deep learning methods, and the proposed deep learning model is fit for post-processing tasks of ERT.\",\"PeriodicalId\":2,\"journal\":{\"name\":\"ACS Applied Bio Materials\",\"volume\":\"20 4\",\"pages\":\"\"},\"PeriodicalIF\":4.6000,\"publicationDate\":\"2024-06-13\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"ACS Applied Bio Materials\",\"FirstCategoryId\":\"5\",\"ListUrlMain\":\"https://doi.org/10.1088/1361-6501/ad57e0\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q2\",\"JCRName\":\"MATERIALS SCIENCE, BIOMATERIALS\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"ACS Applied Bio Materials","FirstCategoryId":"5","ListUrlMain":"https://doi.org/10.1088/1361-6501/ad57e0","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"MATERIALS SCIENCE, BIOMATERIALS","Score":null,"Total":0}
引用次数: 0
摘要
监测工业流程中的多相流分布以优化生产是一项具有挑战性的工作。电阻断层扫描(ERT)可用于观察多相流的内部分布。图像重建在 ERT 中起着至关重要的作用。然而,逆问题的非线性和多拟性使得 ERT 的图像重建成为一项挑战,而先进成像算法的开发在过去一直备受关注。本研究提出了一种改进的 U 型深度学习模型,它结合了 UNet++ 的多尺度特征提取和 Res2Net 的残差特征融合的优点。该网络用于对传统 ERT 图像重建方法的预重建结果进行后处理,将传统方法的泛化能力和深度学习方法的灵活特征提取优势结合起来。为了验证所提模型的泛化能力和有效性,我们设计了仿真和实验。仿真和实验结果表明,所提出的 U 型网络方法优于其他深度学习方法,所提出的深度学习模型适合 ERT 后处理任务。
Res2-UNet++: A Deep Learning Image Post-Processing Method for Electrical Resistance Tomography
It is challenging to monitor the multiphase flow distribution in the industrial processes in order to optimize the production. Electrical resistance tomography (ERT) can be used to visualize the inner distribution of multiphase flow. The image reconstruction plays a vital role in ERT. However, the nonlinearity and ill-posedness of inverse problem make the image reconstruction of ERT a challenge, and the development of advanced imaging algorithm has attracted much attention in the past. In this work, an improved U-shaped deep learning model is proposed, which combines the advantages of multi-scale feature extraction of UNet++ and residual feature fusion of Res2Net. The network is used to post-process the prereconstruction result of traditional ERT image reconstruction methods, where the generalization ability of the traditional methods and the flexible feature extraction advantage of deep learning methods can be combined. Simulations and experiments are designed to verify the generalization ability and the effectiveness of the proposed model. Both simulation and experimental results show that the proposed U-shaped network approach outperforms other deep learning methods, and the proposed deep learning model is fit for post-processing tasks of ERT.
期刊介绍:
ACS Applied Bio Materials is an interdisciplinary journal publishing original research covering all aspects of biomaterials and biointerfaces including and beyond the traditional biosensing, biomedical and therapeutic applications.
The journal is devoted to reports of new and original experimental and theoretical research of an applied nature that integrates knowledge in the areas of materials, engineering, physics, bioscience, and chemistry into important bio applications. The journal is specifically interested in work that addresses the relationship between structure and function and assesses the stability and degradation of materials under relevant environmental and biological conditions.