Skip to content

Provide nnUNet Ensemble Support #8

Description

@SimoneBendazzoli93

Summary

Add support for nnUNet ensemble within MONet-Bundle to enable model ensembling using nnUNet predictions and workflows.

Motivation

Supporting nnUNet ensemble can improve segmentation performance by combining results from multiple models, a method widely adopted in medical imaging challenges. This will increase the flexibility and effectiveness of MONet-Bundle when working with nnUNet-trained models.

Expected Functionality

  • Allow users to run nnUNet ensembles as part of the MONet-Bundle workflow.
  • Provide configuration options for different ensemble schemes (e.g., majority voting, averaging).
  • Documentation and examples on how to enable and use nnUNet ensembling.

Acceptance Criteria

  • A user can configure and trigger an nnUNet ensemble through MONet-Bundle.
  • Proper documentation is provided.
  • Unit and/or integration tests cover the new ensemble workflow.

Additional Context

  • Reference nnUNet's existing ensemble modules for implementation inspiration.
  • Ensure compatibility with MONet-Bundle’s current design.

Metadata

Metadata

Assignees

No one assigned

    Labels

    enhancementNew feature or request

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions