brycewang-stanford WACV Topic Selection
WACV is the applications-first winter member of the Computer Vision Foundation's four-venue calendar (CVPR in June, ICCV/ECCV alternating in autumn, WACV in winter). This skill decides whether to send a paper here — and, if so, which track — before other skills decide how. Facts are the WACV 2026/2027 cycles as read on 2026-07-09.
The core question
Ask two things in order. First: is the contribution about applying computer vision to a real problem, or advancing vision method for its own sake? WACV's name — Applications of Computer Vision — is a genuine editorial stance, and its Applications track exists to review systems that work under real constraints. Second: does the winter, two-round calendar fit your timeline better than waiting for a single-deadline sibling? A paper that is nearly ready in mid-summer can enter WACV Round 1 and, if close, be invited to revise — rather than sitting until the next CVPR cycle.
Routing across the CVF calendar (from WACV's seat)
| You have… | WACV-shaped if… | Route elsewhere if… |
|---|---|---|
| A vision system for a real domain (agriculture, medical, retail, robotics) under real constraints | The constraint (power, data scarcity, in-the-wild robustness) is the contribution | It is a generic method with an arbitrary demo domain → CVPR/ICCV/ECCV |
| A new architecture / algorithm | It advances a vision mechanism and you want the winter slot and the R&R net | It is a field-defining method aimed at the largest audience → CVPR |
| A benchmark / re-evaluation of a deployed task | It corrects how a real task is measured or constrained | It is a broad foundation-model benchmark → a larger venue |
| A 3D / geometry contribution | It is applied and constraint-driven | It is core 3D reconstruction method → 3DV or CVPR |
| A theory / learning result with a vision demo | (rarely WACV) | The claim is statistical/learning-theoretic → NeurIPS/ICML/AISTATS |
WACV is not a "CVPR reject bin." It is a real venue with its own applications identity, a respectable acceptance profile, and a review model designed to improve papers across two rounds rather than gate them once.
Choosing the track
Once WACV is the answer, the track is the next decision, and it changes how you write the first page:
- Applications track — reviewed on systems-level innovation, novelty of the application domain, and comparative assessment under real-world constraints (low-power inference, data scarcity, robustness in the wild). Choose this when the value is that the system works under a named constraint, not that the math is new.
- Algorithms track — reviewed on algorithmic novelty and quantified evaluation against current alternatives. Choose this when you introduce a genuinely new method or architecture and can show matched-baseline wins.
Authors pick one primary track. If a paper seems to fit both, choose the track where the contribution is stronger, and frame the first page to argue for that lens — a mis-tracked paper draws the wrong review question and heads for Revise and Resubmit.
A quick self-test
Is the deployment constraint the point? yes → Applications track, WACV strong
Is a new vision mechanism the point? yes → Algorithms track (WACV) or CVPR/ICCV/ECCV
Is vision only the demo for a learning claim? yes → route to NeurIPS/ICML/AISTATS
Do you want a revision path if you're close? yes → WACV Round 1 fits
Is the claim about a non-vision modality's math? yes → not WACV
Reverify each cycle
- The two-track structure and each track's review criteria.
- WACV's place and timing in the CVF calendar and the two-round dates.
- Whether the winter cycle's deadlines actually beat your next sibling-venue option.
Output format
[Verdict] WACV / route to <CVPR|ICCV|ECCV|3DV|journal|other>
[Why] <one line: applications-under-constraint vs method-for-its-own-sake>
[Track if WACV] Applications / Algorithms — because <criterion match>
[Timing] winter two-round fits: yes/no
[Risk] <the reviewer question this framing must pre-empt>