{"title":"耳鼻喉显微外科解剖结构分割的迁移学习。","authors":"Xin Ding, Yu Huang, Yang Zhao, Xu Tian, Guodong Feng, Zhiqiang Gao","doi":"10.1002/rcs.2634","DOIUrl":null,"url":null,"abstract":"<div>\n \n \n <section>\n \n <h3> Background</h3>\n \n <p>Reducing the annotation burden is an active and meaningful area of artificial intelligence (AI) research.</p>\n </section>\n \n <section>\n \n <h3> Methods</h3>\n \n <p>Multiple datasets for the segmentation of two landmarks were constructed based on 41 257 labelled images and 6 different microsurgical scenarios. These datasets were trained using the multi-stage transfer learning (TL) methodology.</p>\n </section>\n \n <section>\n \n <h3> Results</h3>\n \n <p>The multi-stage TL enhanced segmentation performance over baseline (mIOU 0.6892 vs. 0.8869). Besides, Convolutional Neural Networks (CNNs) achieved a robust performance (mIOU 0.8917 vs. 0.8603) even when the training dataset size was reduced from 90% (30 078 images) to 10% (3342 images). When directly applying the weight from one certain surgical scenario to recognise the same target in images of other scenarios without training, CNNs still obtained an optimal mIOU of 0.6190 ± 0.0789.</p>\n </section>\n \n <section>\n \n <h3> Conclusions</h3>\n \n <p>Model performance can be improved with TL in datasets with reduced size and increased complexity. It is feasible for data-based domain adaptation among different microsurgical fields.</p>\n </section>\n </div>","PeriodicalId":50311,"journal":{"name":"International Journal of Medical Robotics and Computer Assisted Surgery","volume":null,"pages":null},"PeriodicalIF":2.3000,"publicationDate":"2024-05-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Transfer learning for anatomical structure segmentation in otorhinolaryngology microsurgery\",\"authors\":\"Xin Ding, Yu Huang, Yang Zhao, Xu Tian, Guodong Feng, Zhiqiang Gao\",\"doi\":\"10.1002/rcs.2634\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div>\\n \\n \\n <section>\\n \\n <h3> Background</h3>\\n \\n <p>Reducing the annotation burden is an active and meaningful area of artificial intelligence (AI) research.</p>\\n </section>\\n \\n <section>\\n \\n <h3> Methods</h3>\\n \\n <p>Multiple datasets for the segmentation of two landmarks were constructed based on 41 257 labelled images and 6 different microsurgical scenarios. These datasets were trained using the multi-stage transfer learning (TL) methodology.</p>\\n </section>\\n \\n <section>\\n \\n <h3> Results</h3>\\n \\n <p>The multi-stage TL enhanced segmentation performance over baseline (mIOU 0.6892 vs. 0.8869). Besides, Convolutional Neural Networks (CNNs) achieved a robust performance (mIOU 0.8917 vs. 0.8603) even when the training dataset size was reduced from 90% (30 078 images) to 10% (3342 images). When directly applying the weight from one certain surgical scenario to recognise the same target in images of other scenarios without training, CNNs still obtained an optimal mIOU of 0.6190 ± 0.0789.</p>\\n </section>\\n \\n <section>\\n \\n <h3> Conclusions</h3>\\n \\n <p>Model performance can be improved with TL in datasets with reduced size and increased complexity. It is feasible for data-based domain adaptation among different microsurgical fields.</p>\\n </section>\\n </div>\",\"PeriodicalId\":50311,\"journal\":{\"name\":\"International Journal of Medical Robotics and Computer Assisted Surgery\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":2.3000,\"publicationDate\":\"2024-05-20\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"International Journal of Medical Robotics and Computer Assisted Surgery\",\"FirstCategoryId\":\"3\",\"ListUrlMain\":\"https://onlinelibrary.wiley.com/doi/10.1002/rcs.2634\",\"RegionNum\":3,\"RegionCategory\":\"医学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q2\",\"JCRName\":\"SURGERY\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"International Journal of Medical Robotics and Computer Assisted Surgery","FirstCategoryId":"3","ListUrlMain":"https://onlinelibrary.wiley.com/doi/10.1002/rcs.2634","RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"SURGERY","Score":null,"Total":0}
Transfer learning for anatomical structure segmentation in otorhinolaryngology microsurgery
Background
Reducing the annotation burden is an active and meaningful area of artificial intelligence (AI) research.
Methods
Multiple datasets for the segmentation of two landmarks were constructed based on 41 257 labelled images and 6 different microsurgical scenarios. These datasets were trained using the multi-stage transfer learning (TL) methodology.
Results
The multi-stage TL enhanced segmentation performance over baseline (mIOU 0.6892 vs. 0.8869). Besides, Convolutional Neural Networks (CNNs) achieved a robust performance (mIOU 0.8917 vs. 0.8603) even when the training dataset size was reduced from 90% (30 078 images) to 10% (3342 images). When directly applying the weight from one certain surgical scenario to recognise the same target in images of other scenarios without training, CNNs still obtained an optimal mIOU of 0.6190 ± 0.0789.
Conclusions
Model performance can be improved with TL in datasets with reduced size and increased complexity. It is feasible for data-based domain adaptation among different microsurgical fields.
期刊介绍：
The International Journal of Medical Robotics and Computer Assisted Surgery provides a cross-disciplinary platform for presenting the latest developments in robotics and computer assisted technologies for medical applications. The journal publishes cutting-edge papers and expert reviews, complemented by commentaries, correspondence and conference highlights that stimulate discussion and exchange of ideas. Areas of interest include robotic surgery aids and systems, operative planning tools, medical imaging and visualisation, simulation and navigation, virtual reality, intuitive command and control systems, haptics and sensor technologies. In addition to research and surgical planning studies, the journal welcomes papers detailing clinical trials and applications of computer-assisted workflows and robotic systems in neurosurgery, urology, paediatric, orthopaedic, craniofacial, cardiovascular, thoraco-abdominal, musculoskeletal and visceral surgery. Articles providing critical analysis of clinical trials, assessment of the benefits and risks of the application of these technologies, commenting on ease of use, or addressing surgical education and training issues are also encouraged. The journal aims to foster a community that encompasses medical practitioners, researchers, and engineers and computer scientists developing robotic systems and computational tools in academic and commercial environments, with the intention of promoting and developing these exciting areas of medical technology.
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号
