{"title":"基于快速RCNN的多特征提取器乳腺肿瘤检测比较","authors":"Zhen Zhang, Yaping Wang, Jiankang Zhang, X. Mu","doi":"10.1109/ISNE.2019.8896490","DOIUrl":null,"url":null,"abstract":"The Deep learning algorithm shows is strong capability in pattern recognition tasks such as object detection and speech recognition. Comparing with the typical machine learning methods which require extraction of manual features, deep learning algorithm has a more powerful feature learning ability. In this paper, the deep learning associated object detection method is applied to developed to locate and classify lesions for the detection of medical breast masses. At the same time, transfer learning based on the network of Faster RCNN is also introduced. Furthermore, five feature extractors of the network, which are ResNet101, inception V2, inception V3, Mobilenet, and inception ResNet V2, are investigated for exploring the impact for the model. Digital Database for Screening Mammography(DDSM) dataset is used in the investigation, and the performances of the models associated with the five feature extractors are compared separately in detecting benign and malignant breasts, based on the ROC trade-off curves. The simulation results demonstrate that the classification model with Inception ResNet V2 feature extractor exhibit the best performance, compared with the other four feature extractors.","PeriodicalId":405565,"journal":{"name":"2019 8th International Symposium on Next Generation Electronics (ISNE)","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2019-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"11","resultStr":"{\"title\":\"Comparison of multiple feature extractors on Faster RCNN for breast tumor detection\",\"authors\":\"Zhen Zhang, Yaping Wang, Jiankang Zhang, X. Mu\",\"doi\":\"10.1109/ISNE.2019.8896490\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"The Deep learning algorithm shows is strong capability in pattern recognition tasks such as object detection and speech recognition. Comparing with the typical machine learning methods which require extraction of manual features, deep learning algorithm has a more powerful feature learning ability. In this paper, the deep learning associated object detection method is applied to developed to locate and classify lesions for the detection of medical breast masses. At the same time, transfer learning based on the network of Faster RCNN is also introduced. Furthermore, five feature extractors of the network, which are ResNet101, inception V2, inception V3, Mobilenet, and inception ResNet V2, are investigated for exploring the impact for the model. Digital Database for Screening Mammography(DDSM) dataset is used in the investigation, and the performances of the models associated with the five feature extractors are compared separately in detecting benign and malignant breasts, based on the ROC trade-off curves. The simulation results demonstrate that the classification model with Inception ResNet V2 feature extractor exhibit the best performance, compared with the other four feature extractors.\",\"PeriodicalId\":405565,\"journal\":{\"name\":\"2019 8th International Symposium on Next Generation Electronics (ISNE)\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2019-10-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"11\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2019 8th International Symposium on Next Generation Electronics (ISNE)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/ISNE.2019.8896490\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2019 8th International Symposium on Next Generation Electronics (ISNE)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ISNE.2019.8896490","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Comparison of multiple feature extractors on Faster RCNN for breast tumor detection
The Deep learning algorithm shows is strong capability in pattern recognition tasks such as object detection and speech recognition. Comparing with the typical machine learning methods which require extraction of manual features, deep learning algorithm has a more powerful feature learning ability. In this paper, the deep learning associated object detection method is applied to developed to locate and classify lesions for the detection of medical breast masses. At the same time, transfer learning based on the network of Faster RCNN is also introduced. Furthermore, five feature extractors of the network, which are ResNet101, inception V2, inception V3, Mobilenet, and inception ResNet V2, are investigated for exploring the impact for the model. Digital Database for Screening Mammography(DDSM) dataset is used in the investigation, and the performances of the models associated with the five feature extractors are compared separately in detecting benign and malignant breasts, based on the ROC trade-off curves. The simulation results demonstrate that the classification model with Inception ResNet V2 feature extractor exhibit the best performance, compared with the other four feature extractors.