wacv-related-work
GitHub指导撰写WACV论文相关工作部分，涵盖定位贡献、覆盖五大文献维度、处理并发工作、遵守双盲自引规范及验证引用准确性，避免被误认为重复工作或忽视前沿。
Trigger Scenarios
Install
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill wacv-related-work -g -y
SKILL.md
Frontmatter
{
"name": "wacv-related-work",
"description": "Use when writing or auditing a WACV related-work section, covering how to position an applications or algorithms contribution against fast-moving vision literature, sibling-venue (CVPR\/ICCV\/ECCV) and arXiv concurrency, application-domain literature, double-blind self-citation, and verifying that cited \"WACV papers\" are actually WACV papers."
}
WACV Related Work
Use this to position a WACV paper so reviewers see the gap it fills, not a wall of citations. WACV's applications identity adds a shelf most vision venues underweight — the domain literature — and its two-round model means a thin related-work section is a predictable revision request. Facts are the WACV 2026/2027 cycles as read on 2026-07-09.
The shelves to cover
| Shelf | What it establishes | Failure if missing |
|---|---|---|
| Vision method lineage | Where your method sits among CV approaches | Reviewer thinks you reinvented a known method |
| Sibling-venue recent work | Awareness of CVPR/ICCV/ECCV state of the art | Looks unaware of the field's frontier |
| arXiv / concurrent work | Honest handling of near-simultaneous results | Accused of ignoring obvious concurrency |
| Application-domain literature | That you understand the real problem (agriculture, medical, robotics) | Applications reviewer sees a naive domain model |
| Your own prior work (anonymized) | Continuity without breaking double-blind | Identity leak or apparent self-plagiarism |
The domain shelf is what separates a strong WACV applications paper from a generic method with a domain sticker: cite the deployment literature that defines the real constraint, not only the vision papers that share your architecture.
Positioning, not enumeration
Each cited cluster should end in a specific reason it does not already solve your problem — "these run in daylight and lose low-contrast boundaries below 20 lux," not "prior work has limitations." For an Applications-track paper, the gap is usually a constraint no prior system met; for an Algorithms-track paper, it is a mechanism no prior method used.
Concurrency at arXiv speed
For each closely related recent preprint/paper:
- Same problem, different method? → contrast approaches; note it is concurrent.
- Same method, different problem? → clarify what your setting adds.
- Genuinely concurrent (within months)? → cite and distinguish; do not claim it as prior
or ignore it. Reviewers know the arXiv timeline.
Double-blind self-citation
Cite your own prior work in the third person ("Prior work [12] showed…"), never "our previous paper." Keep repository and project-page links out of the submission. A revised Round 2 paper must stay just as anonymous as the Round 1 version — do not let a revision reintroduce an identifying citation.
Verify the venue of every "WACV" citation
A thecvf.com open-access URL is only a WACV citation if its path reads WACV20XX; many
famous vision papers are CVPR/ICCV/ECCV. Before attributing a paper to WACV, confirm the
open-access path segment or the dblp record — misattributing a sibling-venue paper to WACV
is a credibility cost a careful reviewer will notice.
Reverify each cycle
- Any related-work or citation-format expectations in the current guidelines.
- The double-blind and external-link rules for the submission version.
- That cited WACV papers match the WACV proceedings, not a sibling venue.
Output format
[Shelves covered] vision / sibling-venue / concurrent / domain / self — missing: <list>
[Gap statements] each cluster ends in a specific gap: yes/no
[Domain literature] real-problem sources cited: yes/no
[Concurrency] recent preprints cited and distinguished: yes/no
[Anonymity] self-cites third-person, no links: yes/no
[Venue check] all "WACV" cites verified as WACV: yes/no
Version History
- 9f86f09 Current 2026-07-19 17:54
