Workflow understanding
Phase, step, event, and action recognition for structuring long procedure recordings.
Typical labels
Phase labels, step labels, event timestamps, temporal boundaries
Research hub
Datasets, models, benchmarks, and source material for Surgical AI.
This area is for researchers, developers, and technical users who need a structured view of the field. The main site stays focused on learning, demos, community, and The Signal.
Datasets
Video, frames, annotations, procedure labels, and access terms.
Open sectionModels
Research implementations, model families, inputs, tasks, and maturity.
Open sectionResources
Guidelines, challenge hubs, reporting papers, and evaluation references.
Open sectionContribution
How to prepare a dataset, project, or teaching resource for review.
Open sectionVideo tasks
Separate the research question before comparing data or models.
The same operative recording can support different tasks. Mixing them together makes the catalog harder to use and the claims harder to interpret.
Scene understanding
Instrument, anatomy, object, depth, and segmentation tasks describing what is visible in the field.
Typical labels
Boxes, masks, keypoints, depth maps, anatomy labels
Skill and quality signals
Assessment tasks that require explicit definitions, trained reviewers, and careful claims.
Typical labels
Rating scales, safety criteria, motion features, task completion labels
Annotation and benchmarking
Dataset design, inter-rater disagreement, leakage checks, and benchmark selection.
Typical labels
Label protocols, splits, metadata, adjudication rules, evaluation metrics
Datasets
Find data by procedure, task, annotation type, access, and modality.
Use datasets as source material, not as interchangeable benchmarks. Always check the original terms before download or reuse.
Searching catalog...
Models
Compare models by task, input, family, source, and evidence maturity.
The model catalog focuses on what a model was built to do, what data it expects, and what level of evaluation is actually available.
Searching catalog...
Resources
Guidelines, challenge hubs, reporting papers, and evaluation references.
This section is for materials that help researchers design, evaluate, report, or contextualize Surgical AI work.
Searching catalog...
Suggest a resource or prepare a contribution.
Submit official dataset links, model repositories, benchmark hubs, reporting papers, or resource corrections. Data contribution requires governance, consent review, de-identification, and institutional readiness before any upload pathway is opened.
The Weekly Signal
Follow new datasets, models, and research notes.
A concise email with practical explainers, useful papers, dataset updates, tools, and ways to participate.