DCA-MI (2nd Ed.) | Data Curation & Augmentation in Medical Imaging | ECCV 2026 2024 archive

Data Curation
& Augmentation
in Medical Imaging.

An ECCV workshop on the data bottleneck for robust medical imaging AI.

September 8-9, 2026 | ECCV Venue Malmö, Sweden Successor to CVPR 2024
iMED figure showing RGB views, depth maps, aligned point clouds, geometry, and tool labels.
FIG. 01iMED RGB / Depth / Geometry
CLiMB2026 colonoscopy localization and mapping benchmark overview.
FIG. 02CLiMB2026 overview
SurgVU sample frames across multiple surgical tasks.
FIG. 03SurgVU surgical tasks

Invited speakers

Sophia Bano

Sophia Bano

UCL, United Kingdom

Robot vision and scene understanding for minimally invasive surgery.
Lena Maier-Hein

Lena Maier-Hein

DKFZ / Heidelberg University, Germany

Surgical data science, benchmarking, and reproducible evaluation.
José M. M. Montiel

José M. M. Montiel

Universidad de Zaragoza, Spain

Visual SLAM, deformable SLAM for endoscopy, EndoMapper.
Mengya Xu

Mengya Xu

CUHK, Hong Kong

Medical AI across MICCAI, IPCAI, and ICRA.

Advisory board

Sophia BanoAdvisor

Sophia Bano

University College London

Researcher focused on robot vision and scene understanding for minimally invasive surgery.

Lena Maier-HeinAdvisor

Lena Maier-Hein

German Cancer Research Center (DKFZ) / Heidelberg University

Researcher in surgical data science, benchmarking, and reproducible evaluation.

Organizers

Sierra Bonilla

Sierra Bonilla

University College London PhD candidate working on surgical computer vision, 3D reconstruction, medical imaging, and the iMED dataset.

Xiaorui Zhang

Xiaorui Zhang

Intuitive Surgical machine learning engineer organizing SurgVU and building surgical video understanding and high-resolution endoscopic datasets.

Mary Jin

Mary Jin

Intuitive Surgical vision system analyst working on computational imaging and joint optical-software design for surgical imaging systems.

Shuoqi Chen

Shuoqi Chen

Intuitive Surgical machine learning and computer vision engineer developing AI-driven imaging technologies for surgical robotics.

Jingpei Lu

Jingpei Lu

Intuitive Surgical research scientist in computer vision and surgical robotics, with work on robot perception and surgical imaging datasets.

Rogerio Nespolo

Rogerio Nespolo

Intuitive Surgical machine learning engineer focused on surgical data science, real-time guidance, and multimodal skill assessment.

Fengyi Jiang

Fengyi Jiang

Uber AV Labs machine learning engineer focused on autonomous-vehicle perception, real-time systems, data curation, and robust computer vision evaluation.

Areas of focus

Dataset curation, benchmarking, and responsible data practices

Synthetic and simulated data

3D perception and spatial medical imaging

Learning, multimodal data, and clinical translation

Highlighted datasets and challenges

iMED processing figure with RGB, depth, aligned point clouds, geometry, and tool labels.
PRIMARY SPOTLIGHT

iMED

Multi-Endoscope Dataset for 3D Perception

340 sequences~170K timepoints4 views per timepoint2 challenge tasksEx vivoPostmortemLive settingsEndoscopic 3D

iMED2026 is a MICCAI/EndoVis challenge on synchronized multi-endoscope data, with relative pose estimation and deformable novel view synthesis tracks for endoscopic 3D perception.

Challenge site
CLiMB2026 colonoscopy localization and mapping benchmark overview.
FEATURED BENCHMARK

CLiMB2026

Colonoscopy Localization and Mapping Benchmark

ColonoscopyLocalizationMappingUnder development

CLiMB2026 is an unpublished benchmark under development for colonoscopy localization and mapping. Public dataset access, paper details, and benchmark statistics will be added after organizer review.

SurgVU sample frames from multiple surgical tasks.
FEATURED DATASET

SurgVU

Surgical Visual Understanding Dataset Series

840 hours~18M frames280 clips12 tools155 sessions8 tasks

SurgVU focuses on large-scale surgical video understanding. DCA-MI highlights it as a dataset case study for real-world video curation and evaluation.

arXiv:2501.09209

Important dates

15 May 2026 — Call for papers posted

07 Jul 2026 — Paper submission deadline

01 Aug 2026 — Notification of acceptance

15 Aug 2026 — Camera-ready due

08–09 Sep 2026 — Workshop at ECCV 2026

iMED processing figure with RGB, depth, aligned point clouds, geometry, and tool labels.
PRIMARY SPOTLIGHT

iMED

Multi-Endoscope Dataset for 3D Perception

340 sequences~170K timepoints4 views per timepoint2 challenge tasksEx vivoPostmortemLive settingsEndoscopic 3D

iMED2026 is a MICCAI/EndoVis challenge on synchronized multi-endoscope data, with relative pose estimation and deformable novel view synthesis tracks for endoscopic 3D perception.

Challenge site
CLiMB2026 colonoscopy localization and mapping benchmark overview.
FEATURED BENCHMARK

CLiMB2026

Colonoscopy Localization and Mapping Benchmark

ColonoscopyLocalizationMappingUnder development

CLiMB2026 is an unpublished benchmark under development for colonoscopy localization and mapping. Public dataset access, paper details, and benchmark statistics will be added after organizer review.

SurgVU sample frames from multiple surgical tasks.
FEATURED DATASET

SurgVU

Surgical Visual Understanding Dataset Series

840 hours~18M frames280 clips12 tools155 sessions8 tasks

SurgVU focuses on large-scale surgical video understanding. DCA-MI highlights it as a dataset case study for real-world video curation and evaluation.

arXiv:2501.09209

15 May 2026 — Call for papers posted

07 Jul 2026 — Paper submission deadline

01 Aug 2026 — Notification of acceptance

15 Aug 2026 — Camera-ready due

08–09 Sep 2026 — Workshop at ECCV 2026

Submit your workshop paper through the DCA-MI OpenReview portal.

Submit on OpenReview

Submission tracks

Submission deadline

07 Jul 2026, 11:59 PM (AoE)

Calculating…

Original Work Track

Peer-reviewed papers

Full papers in ECCV LNCS format. All submissions are reviewed via a standard double-blind OpenReview workflow by a dedicated program committee. Conflicts of interest are handled per ECCV workshop policies. Accepted papers are included in the workshop program for poster presentation.

Full papers, extended abstracts, and dataset or benchmark papers are welcome.

Published Work Track

Already published work

Papers already published and peer-reviewed at major computer vision, machine learning, medical imaging, robotics, or clinical AI conferences or journals. Submissions are not re-reviewed — the main criteria are topical fit and poster-board availability. Authors submit a link to the paper via the workshop’s OpenReview page.

Topics of interest

  • Acquisition pipelines, annotation strategy, and multi-site dataset construction
  • Pseudo-ground-truth generation, calibration, and quality control
  • Clinically meaningful metrics, splits, leakage analysis, and reproducibility reporting
  • Human-in-the-loop and clinician-grounded evaluation methodologies
  • Privacy, fairness, federated learning, and responsible dataset governance
  • Generative and diffusion-based augmentation for images and video
  • Synthetic data for rare diseases and under-represented populations
  • Physics-based simulation, digital twins, and clinical simulators
  • Sim-to-real adaptation and patient-specific synthetic priors
  • SLAM and structure-from-motion for endoscopy and interventional imaging
  • Neural rendering, Gaussian splatting, and implicit 3D representations
  • Calibration and pose estimation under deformation, smoke, and reflection
  • Dataset and pseudo-GT design for reconstruction benchmarks
  • Self-, semi-, weakly-supervised learning and foundation-model adaptation
  • Domain adaptation, distribution shift, and out-of-distribution detection
  • Multimodal clinical datasets, vision-language models, and workflow understanding
  • Deployment safety, uncertainty quantification, and bench-to-bedside evaluation

Submission details & FAQ

What format and page limit should I use?

Original Work: ECCV 2026 LNCS author kitmax. 14 pages excluding references (figures and tables included). Additional reference-only pages are allowed. Use the 2026 template, not CVPR or older ECCV kits. See the ECCV submission policies.

Published Work: no reformatting — submit the camera-ready version as published at the original venue, plus a link on OpenReview.

Which ECCV policies apply?

Follow the ECCV 2026 submission policies — double-blind anonymization, no author-identifying links, dual-submission rules, and completed OpenReview profiles.

How does review work?

Original Work: double-blind peer review via OpenReview, one round, no rebuttal. Published Work: topical-fit screening only, not re-reviewed.

What happens if my paper is accepted?

Poster presentation at DCA-MI 2026. A subset of Original Work papers may be invited for oral presentation.

What about dataset provenance and ethics?

Document source, consent, licensing, and coverage limits. Submissions with unclear protected-data handling may be desk-rejected.

Sierra Bonilla

Sierra Bonilla

University College London PhD candidate working on surgical computer vision, 3D reconstruction, medical imaging, and the iMED dataset.

Xiaorui Zhang

Xiaorui Zhang

Intuitive Surgical machine learning engineer organizing SurgVU and building surgical video understanding and high-resolution endoscopic datasets.

Mary Jin

Mary Jin

Intuitive Surgical vision system analyst working on computational imaging and joint optical-software design for surgical imaging systems.

Shuoqi Chen

Shuoqi Chen

Intuitive Surgical machine learning and computer vision engineer developing AI-driven imaging technologies for surgical robotics.

Jingpei Lu

Jingpei Lu

Intuitive Surgical research scientist in computer vision and surgical robotics, with work on robot perception and surgical imaging datasets.

Rogerio Nespolo

Rogerio Nespolo

Intuitive Surgical machine learning engineer focused on surgical data science, real-time guidance, and multimodal skill assessment.

Fengyi Jiang

Fengyi Jiang

Uber AV Labs machine learning engineer focused on autonomous-vehicle perception, real-time systems, data curation, and robust computer vision evaluation.

07AAdvisory Board
Sophia Bano Advisor

Sophia Bano

University College London

UCL researcher focused on robot vision and scene understanding for minimally invasive surgery.

Lena Maier-Hein Advisor

Lena Maier-Hein

German Cancer Research Center (DKFZ) / Heidelberg University

DKFZ and Heidelberg University researcher in surgical data science, benchmarking, and reproducible evaluation in medical AI.