Skip to content

Latest commit

 

History

History
72 lines (41 loc) · 2.93 KB

README.md

File metadata and controls

72 lines (41 loc) · 2.93 KB
Raw

Back to Projects List

DeepHeart integration with MONAILabel

Key Investigators (subject to change)

  • Matthew Jolley, MD (Children's Hospital of Philadelphia, Philadelphia, PA, United States)
  • Christian Herz, MS (Children's Hospital of Philadelphia, Philadelphia, PA, United States)
  • Danielle F. Pace, PhD (Martinos Center for Biomedical Imaging, CSAIL, MIT, MGH, HMS, Boston, MA, United States)
  • Andras Lasso, PhD (Laboratory for Percutaneous Surgery, Queen’s University, Canada)
  • John Witt (Children's Hospital of Philadelphia, Philadelphia, PA, United States)
  • Sachidanand Alle (NVIDIA)
  • Prerna Dogra (NVIDIA)
  • Andrés Díaz-Pinto (King's College London, UK)

Project Description

Creation of a MONAILabel app for leaflet segmentation of heart valves in 3D echocardiographic (3DE) images. In particular, we have been developing AI models for segmentation of Tricuspid Valve leaflets in 3DE images of patients with Hypoplastic left heart syndrome (HLHS).

Objective

  1. Creation of a MONAILabel app
  2. Bring own UI elements as FCN will need additional user input

Slide 1

image

Slide 2

image

Slide 3

image

Slide 4

image

Slide 5

image

Slide 6

image

Approach and Plan

  1. Use MONAI framework for replacing most of our custom code to minimize overhead
  2. Create MONAILabel app based on ported code
  3. Create custom UI for additional user inputs

Progress and Next Steps

  1. MONAILabel team created new sample app with support for multi-label segmentation (https://github.com/Project-MONAI/MONAILabel/tree/main/sample-apps/segmentation_liver_and_tumor)
  2. MONAILabel team added option to upload local image to the MONAILabel server
  3. DeepHeart MONAILabel app created for the segmentation of tricuspid valve leaflets from 3DE images

MONAILabel