CVPath+Vision4Neuro
One day, two communities. CVPath brings computer vision to computational and digital pathology; Vision4Neuro brings it to neural decoding, behaviour and clinical neuroscience. Each track keeps its own call for papers and reviewing, and they meet in a joint panel and a combined poster session on what the two fields genuinely share: foundation and multimodal models, robustness under real distribution shift, and evidence of clinical readiness.
When
January 4–5, 2027Workshop days of WACV 2027
Where
Disney SpringsLake Buena Vista, Florida, USA · in person
Papers
8 / 4 / 2 pagesRegular papers archival in the WACV 2027 proceedings
Paper deadline
October 9, 2026Regular papers, 23:59 Anywhere on Earth
The workshop is organised as two self-contained halves under a shared umbrella. Each track has its own call for papers, program committee and reviewing process, so you submit to the community that will review your work best. Everything else — opening, panel, poster session, awards — is shared.
Gigapixel whole-slide images, real domain shift, weak supervision and dense prediction at cellular scale are computer vision problems. With pathology foundation models now openly available, they are open to the whole WACV community, not only to groups with clinical data access.
Reconstructing visual experience from brain signals, aligning vision-language representations with cortex, and reading behaviour from video: the modelling questions are computer vision questions, posed on EEG, MEG, fMRI, fNIRS, ECoG and video.
Tissue slides and neural recordings could hardly be more different as data. The methodological questions, however, line up closely — which is exactly what the joint panel and the combined poster session are for.
Self-supervised pretraining on unlabelled data, vision-language alignment, and generative models are reshaping both fields at once — on gigapixel tissue and on neural signals alike.
Scanner, stain and centre shift in pathology; cross-subject, cross-session and cross-device shift in neural decoding. Both need calibration, open-set awareness and honest evaluation protocols.
What actually counts as evidence that a model is ready to touch patient care? The joint panel puts pathologists, neuroscientists and vision researchers in the same room on that question.
The same timeline applies to both tracks, so authors face one set of dates regardless of where they submit.
| Date | Milestone |
|---|---|
| October 9, 2026 | Regular papers (8 pages, archival) — submission deadline (23:59 AoE) |
| October 30, 2026 | Notification to authors, regular papers |
| November 6, 2026 | Short papers (4 pages) and extended abstracts (2 pages) — submission deadline (23:59 AoE) |
| November 20, 2026 | Notification, short papers and extended abstracts |
| November 20, 2026 | Camera-ready deadline, archival papers (23:59 PT) |
| January 4–5, 2027 | Workshop day at WACV 2027 |
All deadlines are 23:59 Anywhere on Earth unless stated otherwise. The OWL challenge follows its own timeline, on the challenge page.
Both tracks use the same three submission types and the same WACV 2027 paper format, so the choice of track is about audience and reviewers, not about rules.
| Submission type | Length | Proceedings | Presentation |
|---|---|---|---|
| Regular paper | 8 pages (excluding references) | Archival — published in the WACV 2027 Workshop Proceedings (IEEE Xplore and CVF Open Access) | Oral or poster |
| Short paper | 4 pages (excluding references) | Non-archival — not published in the proceedings | Poster, possibly lightning talk |
| Extended abstract | 2 pages (excluding references) | Non-archival — not published in the proceedings | Poster, possibly lightning talk |
An interpretability benchmark for hypothesis-free discovery in prostate cancer histopathology. OWL asks a question ordinary leaderboards do not: when a model predicts biochemical recurrence, does its explanation actually hold up — measured with automatic faithfulness metrics and assessed by expert pathologists, alongside predictive performance.
1,419
Prostatectomy cases from three medical centres
3 axes
Prediction, automatic faithfulness, expert reader study
60 min
Dedicated session with top-team presentations
CVPath runs in the morning, including the OWL challenge session; the joint panel opens the afternoon, right after lunch, so neither community drifts away over the break; Vision4Neuro runs in the afternoon; posters from both tracks share a single session.
Shared opening, CVPath invited talks, CVPath orals and the one-hour OWL challenge session.
Joint panel on robustness, generalization and clinical readiness, with speakers from both communities.
Combined poster session plus Vision4Neuro invited talks and oral session.
Best paper awards for both tracks and shared closing remarks.
Only regular papers. Accepted 8-page regular papers are archival and appear in the WACV 2027 Workshop Proceedings on IEEE Xplore and in the CVF Open Access repository. Short papers (4 pages) and extended abstracts (2 pages) are non-archival: they are presented at the workshop but are not published, so the work remains free for submission elsewhere.
Submit to the community best placed to review your work. CVPath covers computational and digital pathology; Vision4Neuro covers neural decoding, behaviour and clinical neuroscience. Each track has its own program committee, and you select the track when you submit. If a paper genuinely spans both, write to either primary contact and we will agree on the better fit.
No. Each track runs its own reviewing process with its own program committee and criteria. The shared parts of the day are the opening, the panel, the poster session, the awards and the website.
The official WACV 2027 author kit, in LaTeX or Word, the same template as the main conference. Page limits exclude references, and supplementary material does not count towards the limit. Reviewing is double-blind, so submissions must be anonymised.
No. WACV 2027 does not accommodate virtual workshops or remote presentations. At least one author of each accepted contribution must register for the conference and present in person at Disney Springs.
Regular papers must be original and not substantially similar to work under review or published elsewhere, following the WACV dual-submission policy. Because short papers and extended abstracts are non-archival, they may cover work that is concurrently under review at another venue, provided the policy of that venue allows it.
The challenge runs on the Grand Challenge platform, within the CVPath track. Registration and the phase dates are published on the challenge page. Taking part does not require submitting a paper, and top-ranking teams present their methods during the challenge session.
No, but the day is designed to reward it: the joint panel after lunch and the single combined poster session are where the two communities actually meet.
Each track has its own primary contact. For anything concerning the day as a whole, write to either — the two committees coordinate on everything shared.
