Automatic Lung Lesion Detection and Segmentation in CT

Principal Investigator

Name
Lin Yang

Degrees
Ph.D.

Institution
University of Florida

Position Title
Associate Professor

Email
lin.yang@bme.ufl.edu

About this CDAS Project

Study
NLST (Learn more about this study)

Project ID
NLST-377

Initial CDAS Request Approval
Nov 29, 2017

Title
Automatic Lung Lesion Detection and Segmentation in CT

Summary
In this project, we will develop an automatic lung lesion detection and segmentation method using machine learning. We will use the NLST datasets as training/testing sets for various object detectors (including deep learning) and modify the best candidate to achieve promising detection accuracy. After lung lesion detection, we will develop an method for accurate lesion volume segmentation using weakly/interactive supervised learning.

Aims

1. investigate methods for automatic lesion detection,
2. develop automatic lesion segmentation on the detected lesion regions,
3. evaluate our detection and segmentation models using the NLST datasets.

Collaborators

Jinzheng Cai, University of Florida