wacv-submission
GitHub在WACV截稿前两周进行投稿审计，涵盖轮次与赛道选择、8页格式限制、双盲匿名要求、OpenReview档案及补充材料规范，并执行去标识化检查以防直接拒稿。
Trigger Scenarios
Install
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill wacv-submission -g -y
SKILL.md
Frontmatter
{
"name": "wacv-submission",
"description": "Use when auditing a WACV submission before an OpenReview round deadline, covering the Round 1 vs Round 2 choice, the Applications vs Algorithms track field, the 8-page-including-figures limit, double-blind anonymity and OpenReview-profile requirements, the anonymized supplement, dual submission, and desk-reject triage."
}
WACV Submission
Run this audit in the two weeks before a WACV round deadline. The specifics below are the WACV 2026/2027 cycles as read on 2026-07-09; chairs rewrite details each edition, and WACV uniquely runs two rounds, so reopen the current Author Guidelines, Dates page, and the correct OpenReview round group at wacv.thecvf.com before acting on any number here.
First decision: which round, which track
Two fields on the OpenReview form change everything downstream, and both are chosen before you write a word of the audit:
- Round 1 or Round 2. They are different OpenReview groups. Round 1 (reported paper deadline June 26, 2026 for WACV 2027) gives you a rebuttal and the Revise-and-Resubmit safety net — a paper that is close but not done can be invited back rather than killed. Round 2 (reported August 28, 2026) is Accept/Reject with no rebuttal; it is for polished work and for revised Round 1 papers. Submitting a borderline paper straight to Round 2 throws away the mechanism WACV built to help it.
- Applications or Algorithms track. The two tracks are reviewed against different criteria (systems-level innovation and comparative assessment vs algorithmic novelty and matched-baseline evaluation). Pick the track where the contribution is stronger, and make the first page argue for that lens — a mis-tracked paper draws a reviewer asking the wrong question, which is a fast route to Revise and Resubmit.
Format discipline, WACV flavor
The cap is 8 pages including figures and tables, in the WACV style, with extra pages for cited references only. That "including figures" clause makes a qualitative grid or a result table a page-budget line item, not a decoration. Build from the current WACV author kit unmodified; treat any style-file edit, margin nudge, or font substitution as a desk-reject trigger. Every author needs a valid OpenReview profile by the round deadline — an incomplete profile on the advisor added late or the intern who wrote one module is desk-reject exposure for the whole paper.
Anonymity and the anonymized supplement
WACV is double-blind: reviewers should not be able to infer authorship "beyond reasonable doubt" from the paper or the supplement. Anonymize acknowledgements (no co-workers, grant IDs), PDF metadata, self-citations (third person), filenames, and any movie titles or overlays in supplementary videos; do not link to sites that identify the authors. A technical report / arXiv preprint is handled under the current prior-publication rule — check it rather than assuming, and never rely on an external link to carry review content.
# Mechanical sweep on the final build, ten minutes before upload
pdftotext paper.pdf - | grep -nEi 'github\.com|drive\.google|project page|our (prior|previous) work' | head
pdfinfo paper.pdf | grep -Ei 'author|creator|producer' # identity in metadata
grep -rn 'vspace\|\\addtolength\|geometry' main.tex # style-tampering tells
pdftotext paper.pdf - | awk 'END{print NR" lines extracted"}' # sanity: text present, not raster
Desk-reject and triage table
| Finding | Severity at WACV | Recoverable? |
|---|---|---|
| Coauthor without a valid OpenReview profile at the round deadline | Desk reject | Only before that deadline |
| Page 9 contains anything besides references | Rejected without review | Only before the round deadline |
| Style file modified / margins compressed | Rejected without review | Fix and rebuild |
| Identity leak (metadata, acknowledgements, linked repo, video overlay) | Desk reject | Sweep before upload |
| Wrong track for the contribution | Not a reject, but Revise-and-Resubmit risk | Re-frame first page before deadline |
| Same work under review elsewhere in violation of the dual-submission rule | Administrative rejection | Withdraw one venue first |
| Borderline paper sent straight to Round 2 | No rebuttal, no R&R net | Prefer Round 1 next time |
Final-week sequence
- Confirm the round and track fields; they set every other decision.
- Freeze the 8-page body; late edits go to the supplement queue, not the main PDF.
- Rebuild from a clean checkout of the WACV author kit; verify no local style patches leaked in and pages 9+ are references only.
- Run the mechanical sweep above; treat every hit as a blocker.
- Confirm every coauthor's OpenReview profile is valid, and that the OpenReview abstract matches the PDF abstract verbatim.
- Upload, re-download, and read your own submission cold — including checking figures at print zoom, since applications reviewers judge qualitative results visually.
Reverify each cycle
- Both round deadlines and whether the two-round model still holds for the target year.
- The page cap and whether references-only overflow still applies.
- The track definitions and review criteria (Applications vs Algorithms).
- The prior-publication / arXiv rule and the dual-submission window.
- Supplementary size/format caps (not fetchable for 2026 — 待核实).
Output format
[WACV submission] ready / blocked / at-risk
[Round] R1 (rebuttal + R&R net) / R2 (Accept-Reject, no rebuttal) — chosen: <one>
[Track] Applications / Algorithms — first page argues for it: yes/no
[Profiles] all coauthors valid on OpenReview: yes/no
[Format] pages used (of 8 incl. figures/tables) · refs-only overflow ok
[Anonymity/links] clean / violations: <list>
[Fix queue] <ordered blockers>
Version History
- 9f86f09 Current 2026-07-19 17:54
