Vision for Biomedical Discovery (CVPath + Vision4Neuro)
Full-day workshop at IEEE/CVF WACV 2027
Orlando, Florida, USA, 4-5 January 2027
https://cvpath-vision4neuro.github.io/wacv2027/

OVERVIEW

Tissue slides and neural recordings could hardly be more different as data, yet
they raise the same methodological questions: foundation and multimodal models,
generalization under real distribution shift, and what counts as evidence that a
model is ready for clinical use. This full-day workshop brings the two
communities into one room, under a shared umbrella, with two self-contained
tracks, a joint panel and a combined poster session.

Each track keeps its own call for papers, program committee and reviewing
process. Authors select the track when submitting.

TRACK 1 - CVPath: Computer Vision for Computational and Digital Pathology

Topics include, but are not limited to:
- Foundation and self-supervised models for pathology
- Whole-slide image analysis and gigapixel understanding
- Cell- and nuclei-level vision; dense prediction at cellular scale
- Imaging beyond H&E: IHC, multiplex/mIF, multi-stain
- Robustness, domain shift and weak supervision in computational pathology
- Multimodal, generative and vision-language models in computational pathology
- AI interpretability and trustworthy AI in computational pathology
- Reproducibility, benchmarks, deployment and clinical translation

The track hosts the OWL challenge (Observing What Models Learn) on interpretable
AI for hypothesis-free discovery in prostate cancer histopathology, run on the
Grand Challenge platform: https://owl.grand-challenge.org/

TRACK 2 - Vision4Neuro: Neural Decoding and Computer Vision for Behavioral and
Clinical Neuroscience

Topics include, but are not limited to:
- Decoding and reconstruction of visual experience from neural signals
(EEG, MEG, fMRI, fNIRS, ECoG)
- Brain-to-image and brain-to-video generation
- Encoding models and representational alignment between vision-language models
and cortex
- Multimodal neuroimaging and fusion
- Brain-computer interfaces, gaze-driven interfaces and assistive communication
- Vision-based behavioural phenotyping: eye tracking and markerless pose analysis
- Clinical and translational applications in neurodevelopmental and
neurodegenerative conditions
- Robustness in the wild: cross-subject and cross-session transfer, domain shift,
calibration, open-set recognition

SUBMISSIONS

Three submission types, all in the WACV 2027 format, page limits excluding
references:
- Regular papers, 8 pages. Archival, published in the WACV 2027 Workshop
Proceedings on IEEE Xplore and in the CVF Open Access repository.
- Short papers, 4 pages. Non-archival, presented as posters or lightning talks.
- Extended abstracts, 2 pages. Non-archival, presented as posters or lightning
talks.

Reviewing is double-blind. Submissions are handled through OpenReview.
WACV 2027 does not accommodate remote presentations: at least one author of each
accepted contribution must register and present on site.

IMPORTANT DATES

Regular papers (archival) 9 October 2026, 23:59 AoE
Notification, regular papers 30 October 2026
Short papers and extended abstracts 6 November 2026, 23:59 AoE
Camera-ready, archival papers 20 November 2026
Workshop 4-5 January 2027

ORGANIZERS

CVPath: Soumick Chatterjee (Human Technopole, Milan), Francesco Ciompi,
Khrystyna Faryna, Evgenia Martynova (Radboudumc), Francesco Merolla (University
of Molise), Antonio Maria Rinaldi, Cristian Tommasino (University of Naples
Federico II).

Vision4Neuro: Marios Petrov, Adham Atyabi, Terrance E. Boult (University of
Colorado Colorado Springs), Frederick Shic (Seattle Children's Research
Institute and University of Washington).

CONTACT

CVPath: cristian.tommasino@unina.it
Vision4Neuro: mpetrov@uccs.edu