Submissions

R2MI 2026 uses OpenReview to manage submissions, reviews, and decisions.

OpenReview submission portal: TBA (a link will be posted here as soon as it is available)

Submission types

1) Accompanying Reproducibility Papers

Authors of papers accepted to the MICCAI 2026 main conference or any other MICCAI 2026 workshop are invited to submit an accompanying reproducibility paper: a short technical companion tightly linked to the accepted paper.

These submissions should focus on practical, technical information needed to reproduce the results of the main paper, for example:

  • Precise description of the experimental setup and pipeline
  • Environment capture (dependencies, hardware requirements, seeds, determinism notes)
  • Training and evaluation details (data splits, preprocessing/augmentation, hyperparameters)
  • Instructions to run experiments and regenerate key figures/tables
  • Pointers to code, models, containers, and (when possible) data access mechanisms

2) Open Submissions (1-page extended abstracts)

To also include non-main-conference contributions (e.g., tools, infrastructures, checklists, negative results, best practices, failure analyses, and lessons learned), we additionally allow a limited number of open submissions as 1-page extended abstracts.

Open submissions are evaluated under the same reproducibility-oriented criteria.

Review process

Each accompanying reproducibility paper will receive a two-part evaluation:

  • Technical review (at least two reviews): clarity, completeness, and soundness of the reproducibility description
  • Reproducibility check (at least one checker): attempt to replicate key results/figures and verify that instructions run as described

Overall, each submission is handled by at least three evaluators (reviews + reproducibility check).

Acceptance criteria

Acceptance is based on:

  1. The technical quality and usefulness of the submission
  2. The outcome of the reproducibility check
    • effective reproduction of experiments/charts whenever possible, or
    • demonstrated end-to-end executability and transparent reporting of limitations

Reproducibility label

Accepted accompanying reproducibility papers will receive a visible reproducibility label indicating that the submission passed the workshop’s reproducibility assessment.
The label will be displayed on the workshop website and within the published proceedings entry. Authors may also report the label on the repository corresponding to the associated MICCAI paper.

Proceedings

We intend accepted accompanying reproducibility papers (with their reproducibility labels) to be published in the MICCAI Satellite Events joint LNCS proceedings (Springer Nature), subject to workshop selection and publisher requirements.

Presentation and registration

R2MI 2026 is oral-only (no posters).
At least one author of each accepted paper must register and present during the workshop (oral presentation).