Deep Learning-based Lesion Detection and Segmentation in Radiology Images - ON-192

Preferred Disciplines: Computer Science (Masters, PhD or Post-Doc)
Company: Arterys, Inc
Project Length: 8-12 months (2 units)
Desired start date: As soon as possible
Location: Toronto, ON or San Francisco, CA
No. of Positions: 1
Preferences: None

About the Company: 

Arterys is working on applying cloud-based artificial intelligence to create radiology-based algorithms that will impact the lives of millions of people. Much of radiology consists of performing tedious tasks that would greatly benefit from automation, such as segmenting anatomical regions, characterizing lesions and writing reports. Additionally, non-radiologist clinicians who make treatment decisions often do not have the ability to synthesize the myriad components of available clinical information to make the best possible treatment decisions for their patients. Arterys’ goal is to use the latest machine learning technology to solve these problems, help clinicians work more efficiently, and make a dramatic impact on patient outcomes.

Project Description:

Standard radiological workflows for oncology typically involve detecting, measuring, characterizing and tracking potentially cancerous lesions. This project will involve creating Deep Learning and Convolutional Neural Network-based models to automate one or more of those workflow elements. The models will be trained and applied in existing Arterys projects for lung and liver imaging, or potentially in some new projects.

Research Objectives:

  • Deep learning-based detection, segmentation and characterization of lesions
  • Comparison of different model architectures, including tradeoffs in complexity, speed, memory usage and accuracy
  • Modification of existing models from the scientific literature to adapt to radiology workflows and data

Methodology:

  • Model architectures
    • U-Net and similar approaches for segmentation
    • Faster R-CNN and similar approaches for detection
    • ResNet and similar approaches for characterization

Expertise and Skills Needed:

  • Strong background in machine learning and deep learning, particularly convolutional neural networks
  • Experience with Python
  • Fluency with modern deep learning packages, particularly TensorFlow and Keras
  • Experience in production software development and unit testing

For more info or to apply to this applied research position, please

  1. Check your eligibility and find more information about open projects
  2. Interested students need to get the approval from their supervisor and send their CV along with a link to their supervisor’s university webpage by applying through the webform.
Program: