agentsclimarketplace

Eccv workflow

Skill brycewang-stanford/Awesome-Journal-Skills/ECCV-Skills/skills/eccv-workflow

Use when running an ECCV campaign end to end — the even-year calendar from February registration to the September conference, the retargeting triangle to CVPR and ICCV seen from the ECCV seat, where the 2026 cycle stands today, and backward planning a 2028 submission from the March deadline.From its SKILL.md

Install
npx -y skills add brycewang-stanford/Awesome-Journal-Skills --skill eccv-workflow

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

SKILL.md

3.9 KB, 965 tokens by cl100k_base, as published. Nobody here has run it

ECCV Workflow

Use this as the project-management view of an ECCV effort. ECCV happens only in even years, run by ECVA with chairs rotating per edition and no APC — costs are registration and travel; proceedings are Springer LNCS with free ECVA mirrors. The scarce resource is not money but calendar: each decision about ECCV is implicitly a decision about CVPR and ICCV too.

Where the 2026 cycle stands (as of 2026-07-08)

Milestone2026 dateStatus today
Registration / paper / supplementFeb 26 / Mar 5 / Mar 12Closed
Reviews out → rebuttal dueMay 2 → May 11Closed
DecisionsJune 17Released
Camera-readyJune 30, AoE (extended)Just passed
Open-access copies (ECVA/Springer)~4 weeks pre-conferenceExpected ~mid-August
Conference, Malmö (Arena + Malmömässan)Sept 8–13; workshops Sept 8–9Ahead

Accepted teams are in the release-runway phase (eccv-camera-ready, eccv-artifact-evaluation). Rejected teams are at the retargeting fork below with June-dated reviews in hand.

The retargeting triangle from the ECCV seat

The vision big three interleave so that from any ECCV event there is a next move within roughly five months:

ECCV decision (Jun, even yr) --reject--> CVPR deadline (Nov, same yr)
CVPR decision (Feb/Mar, odd yr) --reject--> ICCV deadline (Mar, odd yr)  [tight!]
ICCV decision (Jun/Jul, odd yr) --reject--> CVPR deadline (Nov, odd yr)
CVPR decision (Feb, even yr) --reject--> ECCV deadline (Mar, even yr)   [tight!]

The two "tight" hops are near-simultaneous decision/deadline pairs — teams that anticipate them prepare the resubmission before the decision. Off the triangle: BMVC (~spring deadline) and WACV (~summer) absorb solid work that should not wait; ACCV shares the even-year autumn; IJCV/TPAMI take matured, extended versions. For a paper rejected from ECCV 2026 in June, the default path is a reviews-driven revision into CVPR's November 2026 deadline.

Backward plan for ECCV 2028

All 2028 process details (host city, exact dates, page rules, platform) are unannounced — 待核实 at eccv.ecva.net when posted. Plan against the stable shape (registration late Feb, paper early Mar, conference early autumn):

  • Mar–Sep 2027: project selection; note that ICCV 2027's March deadline competes for the same results — decide which venue gets the work early.
  • Oct–Dec 2027: core method working; falsifier experiment done (eccv-experiments); CVPR 2028's November deadline is the explicit fork.
  • Jan 2028: full experimental program; draft in the (new) author kit.
  • Feb 2028: internal review, anonymity sweep, registration step done the week it opens.
  • Early Mar 2028: paper freeze; supplement week for packaging only.

Roles worth naming on an ECCV team

  • Deadline owner: tracks the CET clock and the three-step chain.
  • Reviewer-duty owner: tracks every co-author's assignments, because 2026-style enforcement couples them to the paper's survival.
  • Release owner: runs the June→September artifact runway.
  • Conference owner: registration (a Springer publication precondition in 2026), Schengen visas, poster/talk logistics for Malmö-like venues.

Output format

[Cycle position] pre-2028 planning / 2026 release-runway / 2026 retargeting
[Next hard date] <milestone, date, clock (CET vs AoE), source>
[Triangle decision] <hold for ECCV / route to CVPR-ICCV-BMVC-WACV / journal>
[Role map] <deadline / reviewer-duty / release / conference -> owner>
[Two-year risk] <what makes this work stale by the next even year>

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.