Selected article for: "classification model and segmentation model"
Author: Shuo Jin; Bo Wang; Haibo Xu; Chuan Luo; Lai Wei; Wei Zhao; Xuexue Hou; Wenshuo Ma; Zhengqing Xu; Zhuozhao Zheng; Wenbo Sun; Lan Lan; Wei Zhang; Xiangdong Mu; Chenxi Shi; Zhongxiao Wang; Jihae Lee; Zijian Jin; Minggui Lin; Hongbo Jin; Liang Zhang; Jun Guo; Benqi Zhao; Zhizhong Ren; Shuhao Wang; Zheng You; Jiahong Dong; Xinghuan Wang; Jianming Wang; Wei Xu
Title: AI-assisted CT imaging analysis for COVID-19 screening: Building and deploying a medical AI system in four weeks
  • Document date: 2020_3_23
    • ID: e6q92shw_11
    Snippet: The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/2020.03. 19.20039354 doi: medRxiv preprint image samples from 5 different hospitals (Table S1 ) with 11 different models of CT equipments (Table S2 ) to increase the model's generalization ability. 3) We developed a three-stage annotation and quality control pipeline, allowing inexperienced data annotators to work with senior radiologists to crea.....
    Document: The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/2020.03. 19.20039354 doi: medRxiv preprint image samples from 5 different hospitals (Table S1 ) with 11 different models of CT equipments (Table S2 ) to increase the model's generalization ability. 3) We developed a three-stage annotation and quality control pipeline, allowing inexperienced data annotators to work with senior radiologists to create accurate annotations, with minimal time investment from the radiologists. 4) We got reasonable training result with only 131 positive cases. Then as more data came in, we retrained the model to improve the model accuracy continuously. 5) To make the diagnosis more intuitive to the radiologists and physicians, in addition to the classification model (output positive / negative predictions), we also introduced a segmentation model that highlighted lesion regions for further examination. 6) The entire tool was delivered as a low-cost, plug-and-play device, so the hospital IT staff could set it up in a self-service way, and we could remotely upgrade the models.

    Search related documents:
    Co phrase search for related documents
    • classification model and generalization ability: 1, 2, 3
    • classification model and lesion region: 1, 2
    • classification model and low cost: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13
    • CT equipment and different hospital: 1
    • different hospital and lesion region: 1
    • different hospital and low cost: 1, 2
    • different model and generalization ability: 1, 2, 3, 4, 5
    • different model and lesion region: 1, 2
    • different model and low cost: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10
    • entire tool and low cost: 1, 2
    • generalization ability and low cost: 1