Writing study: CorrCLIP: Reconstructing Patch Correlations in CLIP for Open-Vocabulary Semantic Segmentation

ICCV overviewCorpus atlasBy contribution typeVisual galleryRules & templateWriting patternsPaper studies

← ICCV 2025 oral writing corpus

Dengke Zhang, Fagui Liu, Quan Tang · Official ICCV 2025 oral record

Observed writing choices, not acceptance rules. This study examines the inspected manuscript version. It does not establish why the paper was accepted or that its presentation choices should be copied unchanged.

Figure, table and section ledger

Editorial contribution type: Architecture & representation (primary).

Layout extraction, not a required paper template. Counts are distinct caption labels; a multi-panel figure counts once. Section spans are estimated from heading positions, including floats and whitespace. Method and limitations.

Canonical PDF

11 physical pages

Before references / appendix

≈ 8.12 page-equivalents

Main figures / tables

4 / 8

Attached appendix figures / tables

0 / 0

Attached appendix span

No boundary detected; not assumed zero

Before references / appendix: ≈ 73.82%References / intervening material: ≈ 26.18%Estimated layout envelopes, not text/artwork area. Unresolved appendix boundaries remain within the undivided post-body material.

Extraction notes: No attached appendix boundary detected. Zero means no recognized caption, not independently verified absence.

Section-by-section lengths and source pages

The title/abstract/front matter precede the first numbered heading. A span ends at the next detected top-level heading, not at the last sentence; end-page whitespace is included. References and statements are listed separately. These are not measurements of text-only or artwork area.

Top-level heading	Starts on PDF page	Estimated span	Share of entire PDF
1. Introduction	1	≈ 1.68 pp.	15.27%
2. Related Work	3	≈ 0.59 pp.	5.36%
3. Method	3	≈ 2.17 pp.	19.73%
4. Experiments	5	≈ 3.02 pp.	27.45%
5. Conclusion	8	≈ 0.22 pp.	2%
6. Acknowledgment	9	≈ 0.08 pp.	0.73%
References	9	≈ 2.88 pp.	26.18%

All detected figure and table locations

Main text

• Figure 1 · p. 1
• Figure 2 · p. 2
• Figure 3 · p. 4
• Table 1 · p. 6
• Figure 4 · p. 7
• Table 2 · p. 7
• Table 3 · p. 7
• Table 4 · p. 8
• Table 5 · p. 8
• Table 6 · p. 8
• Table 7 · p. 8
• Table 8 · p. 8

Attached appendix

No recognized caption labels.

Argument and organization

1. 1 Introduction; 2 Related Work

   Uses controlled patch-interaction diagnostics to motivate a segmentation-specific limitation of CLIP.

2. 3 Method

   Separates correlation scope, correlation values, feature refinement and final map correction.

3. 4 Experiments; 5 Conclusion

   Crosses CLIP sizes and benchmark definitions before testing component effects, mask resolution and model variants.

Problem, claim and contribution

The motivating experiment varies inter-class interactions before proposing restrictions, connecting an observed failure pattern to the method design.

Figure 2 and introduction, physical p.2 · Main paper

Argument and section order

The method names two distinct operations on correlation structure, then adds refinement and correction as separate subsections.

Sections 3.2–3.5, pp.4–5 · Main paper

Figures, captions and visual explanation

The overview makes the external mask generator and feature extractor explicit, with scope reconstruction constraining and value reconstruction populating the similarity matrix.

Figure 3, p.4 · Main paper

Tables, comparisons and evidence

Eight benchmarks include related background/no-background variants, while the ablations expose mask-resolution costs and individual branch non-wins.

Section 4.1, p.5; Tables 4 and 8, p.8 · Main paper

Appendix and reproducibility

Computational costs, alternate mask generators, further diagnostics and training-based comparisons are assigned to supplementary material, not inspected.

Sections 4.2–4.4, pp.6–7 · Main paper

Clarity and writing conventions

The method acknowledges that merged masks can contain multiple classes, and the ablation discussion explains degradation from individual refinement branches.

Section 3.2, p.4; Section 4.4, p.8 · Main paper

What to reuse—and what not to infer

Transferable lesson: Distinguish which interactions are allowed from how strongly they are weighted, and connect each named component to a diagnostic or ablation.

Boundary: Training-free inference still depends on pretrained CLIP, DINO and SAM2; it is not CLIP-only or supervision-free development. Eight benchmark scores are not eight independent data sources. Oracle class/object diagnostic masks are motivation rather than deployable inputs, and the full system’s compute is not established by the main-paper accuracy table.

Apply this lesson using the argument and evidence worksheet.

Inspection record and sources

Checked

2026-10-07

Reading status

Writing-focused manuscript scan + selected visuals inspected

Main text

Writing-oriented scan across the main manuscript, physical PDF pp.1–8, covering framing, section progression, method explanations, captions, comparison protocols and concluding scope; reference boundary checked. Selected full-page rasters were inspected. Not exhaustive line-by-line reading, proof verification or experimental reproduction.

Appendix / supplement

Main-paper appendix/supplement references were checked. A separate supplementary source is linked in the inventory but was not retrieved or inspected; its contents are not included in the reading or visual-page counts.

Visually inspected PDF pages

4, 8

This is a writing-focused scan of the manuscript and the stated supplementary scope: framing, organization, visual explanation, evidence presentation and wording. It is not a line-by-line proof audit, reproduction of results or exhaustive scientific review.

• Official CVF open-access ICCV 2025 accepted-paper PDF, retrieved 7 October 2026; CVF states identical to accepted version except watermark; not assumed identical to initial submission · 11 pages