CVPR 2026¶
IEEE/CVF Conference on Computer Vision and Pattern Recognition — Colorado Convention Center, Denver, Colorado, USA, June 3–7, 2026 (Workshops + Tutorials Jun 3–4; main conference Jun 5–7). Expected attendance ~10,000.
Status: Scaffold — meeting is ~3.5 weeks out as of May 11, 2026. Final paper decisions landed Feb 20, 2026; the accepted-paper list and per-paper OpenReview pages are public; camera-ready and the full virtual program are in flight through May. The workshop slate is finalized (Wed Jun 3 / Thu Jun 4 / split half- and full-day formats). This is a light build, not full coverage — see "Scope" below.
Why this is in the vault¶
CVPR is the premier computer-vision venue worldwide (~10,000 attendees, sibling in scale to NeurIPS / ICML) and like those two the overwhelming majority of its program is general-purpose vision — scene understanding, 3D reconstruction, autonomous-driving perception, generative image / video, robotics — with no direct cancer-research relevance. We are not trying to cover it like ISMB or MICCAI. CVPR lives in this vault for one specific reason: it is the largest single feeder of methodology for medical-imaging and pathology foundation models that later appear in AACR / RSNA / MICCAI translational dossiers.
The convergence points:
- Pathology foundation models. The AACR vault carries standalone dossiers for CHIEF, Prov-GigaPath, and UNI. The base architectures (DINOv2 / DINOv3-class tile encoders, slide-level aggregation transformers, multimodal CLIP-style alignment) are CVPR / ICCV / ECCV-native work. When a pathology FM cites "DINOv2-pretrained ViT-L," that ViT-L paper lived at CVPR or NeurIPS, not in Nature Medicine.
- Medical vision-language models. Radiology report generation (CXR-LLaVA, MAIRA, Merlin, RadFM-class), pathology VLMs (CONCH, PathChat, MUSK), and the broader CLIP / SigLIP / LLaVA stack all surface their core architectures at CVPR before clinical translation at RSNA / MICCAI.
- Generative medical imaging. Diffusion priors for MRI/CT reconstruction, synthetic data for fairness audits, and counterfactual generation for explainability all build on CVPR-mainstream diffusion (latent diffusion, ControlNet, DiT, flow matching) work.
- Segmentation methods. SAM / SAM-2 / EfficientSAM / SAM-Med-class adapters dominate medical segmentation downstream, and the parent SAM line is CVPR / Meta work that MICCAI workshops adapt.
- Major feeder for AACR / MICCAI / RSNA. A CVPR June paper on a new ViT pretraining recipe typically appears as a MICCAI September methods paper on pathology-FM v2 and an AACR April translational poster in the following spring — the 10-month adapt-to-oncology arc.
Scope (what this vault covers — and what it does not)¶
In scope (light build):
- The medical-imaging / biomedicine / clinical-translation workshops at CVPR 2026 (six identified; see roster below).
- Main-track papers that match a medical-imaging / pathology / radiology / biomedical-VLM / generative-medical-imaging / segmentation-for-medicine keyword filter (anticipate a few dozen out of ~2,500 main-track papers).
- Architectural building blocks that are known feeders to pathology / radiology FMs even when the parent paper has no medical framing — DINOv3-class self-supervised pretraining recipes, slide-level / multi-instance aggregation transformers, SAM-line segmentation foundation models, frontier multimodal alignment methods (next-gen CLIP / SigLIP).
Out of scope:
- General scene understanding, 3D reconstruction, NeRF / Gaussian splatting (unless explicit medical use), autonomous-driving perception, robotics, video generation for entertainment, face / person reID, AR/VR — the >90% bulk.
- The full ~2,500-paper main proceedings; bulk extraction is explicitly not planned.
- Tutorials and demos outside the medical niches.
- Pure efficiency / inference-acceleration work without a medical downstream.
Anticipated bio / health-relevant workshops¶
The CVPR 2026 workshop slate is finalized; six identified medical / biomedical / clinical workshops (subject to room and schedule changes through May):
| Workshop | Theme | Date | Format |
|---|---|---|---|
| MCV — Medical Computer Vision | 12th iteration; broad medical-imaging methods + applications | Wed Jun 3 | Full day |
| FMV — Foundation Models for Medical Vision | 3rd iteration; medical-imaging FMs, SAM-Med-class, RadFM-class | Wed Jun 3 | Half day |
| MMFM-BIOMED — Multimodal Foundation Models for Biomedicine | 2nd iteration; cross-modal medical FMs, pathology + radiology VLMs | Wed Jun 3 | Half day |
| PHAROS AI Factory for Medical Imaging & Healthcare | EU AI Factory project; medical imaging + healthcare deployment | Wed Jun 3 | Half day |
| Bridging AI and Medical Reality (CV4Clinical) | Real-world clinical translation, reliability, reproducibility | Thu Jun 4 | Half day |
| Med-Reasoner — Medical Reasoning with Vision Language Foundation Models | Reasoning capabilities for medical VLMs; clinician + CV researcher cross | Thu Jun 4 | Half day |
This is a notably heavier medical workshop footprint than ICML (which has 2–3 bio workshops out of ~10) — CVPR's pathology / radiology FM gravity is the explanation.
What we have / will have to work with¶
| Source | Coverage | Notes |
|---|---|---|
| Main proceedings | ~2,500 accepted papers | OpenReview venue CVPR.thecvf.com/2026/Conference; CVF Open Access posts camera-ready PDFs around the meeting. Full open-access proceedings (no paywall, unlike Springer LNCS for MICCAI). |
| Workshop pages | one per workshop, varying maturity | Six medical workshops with public sites; accepted-paper lists drop May–early June. |
| CVF Open Access | post-meeting full-text repository | openaccess.thecvf.com — full PDFs typically up within ~1 week of meeting. |
| arXiv mirrors | most accepted papers preprint | Title-search "CVPR 2026" on arXiv finds the bulk of accepted work; widely cross-posted. |
| bioRxiv / medRxiv mirrors | medical-imaging subset | Pathology / radiology FM papers often dual-post; benchmark on TCGA / CPTAC / MIMIC-CXR cohorts. |
| Social / press | real-time reactions | #CVPR2026 on Bluesky / X / Mastodon; useful for catching which medical workshop talks get traction. |
| Recorded talks | YouTube + CVF | Orals are typically recorded; workshop recording is workshop-dependent. |
| Cross-link with MICCAI 2026 | author + tool overlap is the rule | The same lab that ships a DINOv3-medical at CVPR June will ship a pathology-FM v2 at MICCAI September — single tool entry should consolidate. |
Organization¶
conferences/cvpr-2026/
├── index.md # this page
├── program-notes.md # pre-meeting prep — dates, filter criteria, expected themes, workshop slate
└── tools/ # per-paper / per-tool entries, populated post-meeting
└── index.md # template + scope-filter note
No digest/, no themes.md, no keynotes/, no timeline.md, no mkdocs.yml. The light-build pattern mirrors ICML 2026 and NeurIPS 2026: only tools/ gets populated, and only with the filtered subset.
What's different from full-conference vaults¶
Compared to MICCAI 2026 — the methodologically nearest medical-imaging sibling, with full proceedings coverage planned — CVPR 2026 is scope-limited by design:
- No per-session digests, no per-day digests, no themes synthesis. The general-vision bulk is irrelevant.
- No keynote / plenary coverage unless the speaker is a known medical-imaging / pathology-FM author.
- Tool entries are filtered down to a few dozen, not exhaustive. The filter is keyed on AACR axes + the pathology-FM dossier chain, not on CVPR's own taxonomy.
- Most CVPR 2026 tool entries should explicitly cross-link to a MICCAI 2026 cousin paper (same lab, September version) — the half-year arc is informative.
- Architectural-building-block entries (DINOv3 successor, SAM-3, next-gen CLIP) get an entry only if they have a documented or near-certain medical adoption path. Pure-vision flagship papers without a medical downstream stay out.
Key dates¶
| Date | Event |
|---|---|
| Nov 3, 2025 | Workshop submission deadline — past |
| Nov 7, 2025 | Abstract submission deadline (AoE) — past |
| Nov 13, 2025 | Full paper submission deadline (AoE) — past |
| Nov 20, 2025 | Supplementary materials deadline — past |
| Dec 15, 2025 | Tutorial submission deadline — past |
| Jan 22, 2026 | Reviews released; rebuttal period opens — past |
| Jan 29, 2026 | Rebuttal period closes — past |
| Feb 20, 2026 | Final paper decisions announced — past |
| Mar 8, 2026 | Art submissions deadline — past |
| Mar 20, 2026 | Demo submissions deadline — past |
| Mar 31, 2026 | Broadening Participation deadline — past |
| Apr 23, 2026 | Early registration deadline — past |
| May 13, 2026 | Cancellation deadline — this week |
| Jun 3, 2026 | Workshops + Tutorials Day 1 |
| Jun 4, 2026 | Workshops + Tutorials Day 2 |
| Jun 5–7, 2026 | Main conference (Denver, Colorado Convention Center) |
Sources¶
- CVPR 2026 home
- CVPR 2026 Important Dates
- CVPR 2026 Workshops list
- CVPR 2026 Call for Workshops
- FMV — Foundation Models for Medical Vision (CVPR 2026)
- Med-Reasoner workshop (CVPR 2026)
- CV4Clinical 2026 — Bridging AI and Medical Reality
- MMFM-BIOMED workshop
- CVF Open Access
- CVPR-MIA paper tracker (community-maintained)
Cross-vault links¶
- AACR 2026 — pathology-FM and translational-cancer side. Direct dossier links from CVPR tool pages back to:
- CHIEF — pathology FM, Mahmood Lab
- Prov-GigaPath — pathology FM, Microsoft / Providence
- UNI — pathology FM, Mahmood Lab
- MICCAI 2026 — direct methodological sibling; same labs, half-year offset. Most CVPR medical-imaging tool entries here should cross-link to a MICCAI 2026 cousin.
- RSNA 2026 — clinical-deployment / vendor side of medical-imaging AI. CVPR architecture → MICCAI methods paper → RSNA clinical-trial readout is the canonical 18-month arc.
- NeurIPS 2026 — general ML / FM side; cross-pollinates with CVPR on diffusion methods, multimodal alignment, and biomedicine workshops (LMRL).
- ICML 2026 — sibling light-build template; same scope-filter approach.
Next step¶
- Now (May 11, 2026, T-3.5 weeks): stand up template + program-notes (this commit). Begin scanning the accepted-paper list with the keyword filter; pre-stub the highest-signal entries (pathology FM updates, SAM-line medical adapters, radiology VLM new entrants).
- Pre-meeting (T-1 to T-3 weeks): workshop accepted-paper lists post; pull DOIs and arXiv IDs into stub pages. Cross-check authorship against the MICCAI 2026 early-accept list (which posted May 7) for cousin-paper pairing.
- Live meeting (Jun 3–7): opportunistic capture of slide deposits, GitHub repos, demo links from social. The Wed Jun 3 medical-workshop cluster (MCV / FMV / MMFM-BIOMED / PHAROS) is the highest-signal day.
- Post-meeting (mid-Jun 2026 onward): CVF Open Access proceedings post ~1 week after meeting. Bulk-populate
tools/with the filtered subset. Estimated ~25–50 tool entries. Cross-link into AACR 2026 pathology-FM dossiers and stage MICCAI 2026 (Sep 27–Oct 1) cousin pairings.