Pulmonary nodule detection with deep learning
Principal Investigator
Name
Ji Park
Degrees
M.D
Institution
JLK-INSPECTION
Position Title
Researcher
Email
About this CDAS Project
Study
NLST
(Learn more about this study)
Project ID
NLST-551
Initial CDAS Request Approval
Aug 9, 2019
Title
Pulmonary nodule detection with deep learning
Summary
We aim to develop deep learning network for pulmonary nodule detection and lung nodule malignancy prediction.
Firstly, we intend to train a large number of CT Scans so that we could attempt to achieve high performances on our nodule detection model and then to build malignancy classification model with high prediction probability.
To evaluate our methods, we validate our models on test datasets whether it achieves solid performance results.
The algorithm is expected to achieve high performances on 1) nodule detection and 2) benign/malignancy classification
Firstly, we intend to train a large number of CT Scans so that we could attempt to achieve high performances on our nodule detection model and then to build malignancy classification model with high prediction probability.
To evaluate our methods, we validate our models on test datasets whether it achieves solid performance results.
The algorithm is expected to achieve high performances on 1) nodule detection and 2) benign/malignancy classification
Aims
To build a deep learning model to detect pulmonary nodule with NLST dataset
To train an efficient convolutional network model for nodule malignancy prediction
To predict whether a region of interest is a nodule or non-nodule with our methods
To achieve high nodule malignancy accuracy
Collaborators
None