Yuru Chen, Jing Feng, Juan Liu, Baochuan Pang, Defa Cao, Cheng Li
{"title":"Detection and Classification of Lung Cancer Cells Using Swin Transformer","authors":"Yuru Chen, Jing Feng, Juan Liu, Baochuan Pang, Defa Cao, Cheng Li","doi":"10.4236/jct.2022.137041","DOIUrl":null,"url":null,"abstract":"Lung cancer is one of the greatest threats to human health. It is a very effective way to detect lung cancer by pathological pictures of lung cancer cells. Therefore, improving the accuracy and stability of diagnosis is very impor-tant. In this study, we develop an automatic detection scheme for lung cancer cells based on convolutional neural networks and Swin Transformer. Microscopic images of patients’ lung cells are first segmented using a Mask R-CNN-based network, resulting in a separate image for each cell. Part of the background information is preserved by Gaussian blurring of surrounding cells, while the target cells are highlighted. The classification model based on Swin Transformer not only reduces the computation but also achieves better results than the classical CNN model, ResNet50. The final results show that the accuracy of the method proposed in this paper reaches 96.16%. Therefore, this method is helpful for the detection and classification of lung cancer cells.","PeriodicalId":66197,"journal":{"name":"癌症治疗(英文)","volume":"123 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2022-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"3","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"癌症治疗(英文)","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.4236/jct.2022.137041","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 3
Abstract
Lung cancer is one of the greatest threats to human health. It is a very effective way to detect lung cancer by pathological pictures of lung cancer cells. Therefore, improving the accuracy and stability of diagnosis is very impor-tant. In this study, we develop an automatic detection scheme for lung cancer cells based on convolutional neural networks and Swin Transformer. Microscopic images of patients’ lung cells are first segmented using a Mask R-CNN-based network, resulting in a separate image for each cell. Part of the background information is preserved by Gaussian blurring of surrounding cells, while the target cells are highlighted. The classification model based on Swin Transformer not only reduces the computation but also achieves better results than the classical CNN model, ResNet50. The final results show that the accuracy of the method proposed in this paper reaches 96.16%. Therefore, this method is helpful for the detection and classification of lung cancer cells.