ECCV Submission

Use this to run the final audit on an ECCV upload. The 2026 window has closed (paper deadline March 5, 2026), so treat every number below as a 2026 anchor for post-mortems and 2028 planning, and reopen the next cycle's Dates and Submission Policies pages before quoting anything as current.

The 2026 deadline chain (verified 2026-07-08)

ECCV runs a two-step front door with a trailing supplement — three separate cutoffs, and the clock was CET, not AoE:

Aggregators disagree by one day on the paper deadline (March 5 vs March 6) because of timezone conversion — another reason to read the official Dates page in CET.

Format gate

• LNCS single-column, 14 pages including figures and tables; pages beyond 14 may contain only cited references.
• The 2026 author kit stated the template changed after 2024 — a recycled ECCV 2024 .cls is a compliance risk, not a convenience.
• Single-column LNCS absorbs figures very differently from a two-column draft: full-width figures are cheap, side-by-side panels are not. Budget pages in the real template from the first draft.
• Submission platform: OpenReview, group thecvf.com/ECCV/2026/Conference.

Policy gate

• Dual submission: ECCV 2026 defined a "publication" as any peer-reviewed, accepted written work longer than four pages excluding references; arXiv preprints and technical reports did not count. Violations meant rejection and a report to the other venue.
• arXiv: posting is not an anonymity violation, but the submission must not cite the authors' own public codebase.
• Embargo: papers publicly identified as ECCV submissions during the review embargo faced summary rejection — the media rule bans naming the venue, not discussing the work.
• Reviewer duty: all authors were expected to be willing to review; an author whose assigned reviews were late or "highly irresponsible" (including LLM-written) risked desk rejection of all their submissions.

Pre-upload sweep

□ Registration record exists for this exact title/author set (Feb deadline)
□ PDF built from the current-cycle author kit, 14 pages + refs only
□ CET deadline entered in the team calendar, not AoE
□ No self-identifying codebase links, acknowledgements, or grant strings
□ Dual-submission attestation true under the 4-page definition
□ Nobody on the author list has an unresolved reviewing assignment
□ Supplement plan for the +7-day deadline drafted before the paper freeze

Severity ladder

The last row is the distinctly ECCV-2026 hazard: your submission's fate is coupled to your co-authors' reviewing behavior, so confirm each co-author's reviewer status during the audit, not after decisions.

Missed-window arithmetic

Because ECCV is biennial, a slipped March deadline costs two years at this venue. The realistic reroutes from a March miss: BMVC (spring deadline), WACV (summer deadline), CVPR (November deadline), ICCV (next odd-year March). Route with eccv-workflow rather than rushing a broken upload.

Output format

[ECCV upload verdict] go / fix first / reroute
[Deadline-chain status] <registration / paper / supplement, with CET times>
[Format violations] <page 15 content, template vintage, column abuse>
[Policy violations] <dual-submission / embargo / anonymity / reviewer duty>
[Ordered fixes] <what to change before the CET cutoff>