What the project gives a department
A ready research platform
Supervising student research often stalls on data: it has to be collected, permitted, annotated and checked. In this project that work is done or under way:
- open surgical datasets with checked licences;
- the CVAT annotation tool, deployed for the project;
- an annotation guideline approved by a surgeon;
- a quality-control method with clear metrics.
What is left for you is the reason you supervise: the question, the method and the discussion of results.
Teaching material
Annotated recordings of operations are visual aids for topographic anatomy, operative surgery and general surgery. Structures are outlined and labelled on real laparoscopic images, and the phases of the operation are marked on a timeline. In class you can walk through a cholecystectomy from trocar placement to gallbladder extraction and show what the critical view of safety looks like.
Publications for the department
Joint papers and abstracts on open data, with a reproducible method. Educators who supervise students and review annotation become co-authors under the contribution agreement.
An expert role
Experienced surgeons and anatomists can refine class definitions, resolve disputed frames and review the annotation guideline. That is the clinical track.
International context
Research groups around the world use the project’s datasets. Your students’ results can be compared with published work, which is a good school of scientific honesty.
A semester in the project: a sample plan for a research group
| Weeks | What happens | Outcome |
|---|---|---|
| 1–2 | How medical AI works, data ethics and licences, the annotation guideline | CVAT accounts, guideline read |
| 3–4 | Calibration on frames with a reference annotation | A report on each student’s match with the reference |
| 5–10 | Real annotation, peer review, disputed frames discussed with a surgeon | Accepted annotation hours |
| 7–12 | A small study on one of the topics | Analysis, figures, limitations |
| 13–16 | Writing an abstract and a talk | An abstract for a student conference |
The plan can be compressed into a 4–6 week module or stretched over an academic year that ends with a thesis.
Ways to work together
- A student research group based on the project: three or more students under your supervision.
- Term papers and theses from the list of topics or your own ideas.
- Joint work with computer science students: the medical student owns the clinical meaning, the computer science student owns the model and the code. Paired topics are in the list.
- An elective course, “Artificial intelligence in surgery”, with hands-on annotation; we can design the syllabus together.
- A university partnership: a cooperation agreement. See the partnerships track. A university’s name appears on the website only with its written permission.
What the project takes on
- CVAT accounts and training in the tool;
- ready tasks from checked, de-identified data;
- quality metrics for each student: match with the reference and agreement with other annotators;
- clinical review of disputed cases;
- advice on methods, statistics and analysis code.
What to agree in advance
- Ethics. The project obtains a separate written opinion for annotating open de-identified data. If a study involves surveys, teaching experiments or analysis of the students’ own performance, it needs approval from your university’s ethics committee and the participants’ consent.
- Licences. Most datasets allow non-commercial use only. Teaching and research fit within that. See the licences page.
- Credit. Only the university can count work in the project as a placement or as research work. On request, the project confirms the amount and quality of each student’s work.
How to start
Write to support@opensurgicalvision.org with the subject “Department: cooperation”. Tell us the university, the department, roughly how many students and which format interests you. We will reply and suggest a time for a short online meeting.
Students who want to take part on their own have their own page. Educators at technical universities: see the section on the page for computer science students.