NA-MIC Project Weeks

Back to Projects List

DeCA: Dense Correspondence Analysis Toolkit for Shape Analysis

Key Investigators

• Sara Rolfe (Seattle Children's, USA)
• Murat Maga (Seattle Children's, USA)

Project Description

DeCA (Dense Correspondence Analysis) is an open-source tool for biologists and other researchers using 3D imaging. DeCA integrates biological insights in the form of homologous landmark points with dense surface registration to provide highly detailed shape analysis of smooth and complex structures that are typically challenging to analyze with sparse manual landmarks alone.

Currently, DeCA exists as a prototype that can be run within 3D Slicer. We have collected preliminary feedback from initial users to improve the interface and workflow. The goal of this project is make and test these updates and publish DeCA as an extension.

Objective

1. Objective A. Implement an improved workflow for the DeCA module
2. Objective B. Publish DeCA extension

Approach and Plan

1. Update the DeCA interface to simplify running analysis
2. Demo/test improvements and collect feedback
3. Document workflow
4. Publish extension

Progress and Next Steps

1. Document changes from initial user testing
2. Meetings to discuss workflow/interface and future applications to new datasets
3. Completion of user interface changes with two fully automated workflows
4. Updates to GUI support and logic functions to support new workflow
5. Landmark subsetting function added to DeCAL
6. Testing of DeCAL dense landmarking

Illustrations

Background and References

Source: https://github.com/smrolfe/DeCA

Publications:

• Rolfe, S. M., and A. Murat Maga. “DeCA: A Dense Correspondence Analysis Toolkit for Shape Analysis.” International Workshop on Shape in Medical Imaging. Cham: Springer Nature Switzerland, 2023.

• Rolfe, S. M., Mao, D., & Maga, A. M. (2024). Streamlining Asymmetry Quantification in Fetal Mouse Imaging: A Semi-Automated Pipeline Supported by Expert Guidance. bioRxiv, 2024-10.