Research on the evolutionary algorithm of CT images and pathological images based on multiple time scales
Principal Investigator
Name
Lijing Sun
Degrees
Master of Philosophy
Institution
Nanjing Tech University
Position Title
Student
Email
About this CDAS Project
Study
NLST
(Learn more about this study)
Project ID
NLST-1250
Initial CDAS Request Approval
May 16, 2024
Title
Research on the evolutionary algorithm of CT images and pathological images based on multiple time scales
Summary
Aiming at the malignant risk assessment and evolution prediction of lung nodules, the dynamic evolution prediction method of lung nodules based on malignant risk assessment is studied. Through the time incremental lung nodule difference information, and the related follow-up data, the lung nodules are classified, and the dynamic evolution prediction model based on multimodal information is constructed, which solves the problem of predicting the evolution of the lung nodules with unknown malignancy degree. I would like to have access to the imaging and pathology images of the dataset and would also like to be able to apply it to follow-up data (data related to survival time).
Aims
1. use deep learning to extract pathology image features at different levels, including cells and tissues
2. extraction of CT images and associated multimodal learning using pre-trained models.
3. use a tumour prediction model based on follow-up data to predict the benign and malignant nature of lung nodules as well as pathological information.
4. convergence of results using multiple loss functions.
Collaborators
none