Lung nodule characterization using machine learning techniques
Principal Investigator
Name
Jack Lin
Degrees
M.S.
Institution
N/A
Position Title
independent researcher
Email
About this CDAS Project
Study
NLST
(Learn more about this study)
Project ID
NLST-270
Initial CDAS Request Approval
Jan 4, 2017
Title
Lung nodule characterization using machine learning techniques
Summary
We are interested in extracting features of lung nodules via convolutional neural network and building predictive models using those features.
Primarily, we will use the corpora to build a lung nodule detection model.
Secondarily, we will validate a lung nodule detection model previously developed on a small training data set against NLST data set. It is well known that greater data variability associated with large training data sets help improve generalisability. As such, we will also compare the generalisability of model developed on NLST against a separate test data sets.
Lung cancer is the leading cause of cancer-related death in China and will likely remain so for the foreseeable future due to its air pollution and industrial development. We aim to tackle this problem by developing detection and predictive models.
Primarily, we will use the corpora to build a lung nodule detection model.
Secondarily, we will validate a lung nodule detection model previously developed on a small training data set against NLST data set. It is well known that greater data variability associated with large training data sets help improve generalisability. As such, we will also compare the generalisability of model developed on NLST against a separate test data sets.
Lung cancer is the leading cause of cancer-related death in China and will likely remain so for the foreseeable future due to its air pollution and industrial development. We aim to tackle this problem by developing detection and predictive models.
Aims
We will develop a lung nodule detection model and a predictive diagnostic model.
Collaborators
N/A