(Learn more about this study)
Initial CDAS Request Approval
Aug 18, 2020
Learning with Sure data for Lung Cancer Prediction
In certain cases (e.g. lung nodule prediction), ground truth labels manually annotated by radiologists (unsure data) are often based on subjective assessment, which lack pathological-proven benchmarks (sure data) at the nodule-level. To address this issue, we would like to use the NLST Dataset with sure data as a strong supervision for training and domain adaptation.
1. Under the supervision of NLST sure dataset, we want to objectively reveal the hidden drawbacks of unsure data served for lung cancer prediction.
2. Investigate the domain adaptation between the sure and unsure datasets.
3. Research on the dynamic auto-assignment strategy of radiologists' voting scores with feedback of sure data model.
Prof Guang-Zhong Yang, SJTU
Dr Yun Gu, SJTU
Mr Yulei Qin, SJTU