Writing study: E-SAM: Training-Free Segment Every Entity Model

ICCV overviewCorpus atlasBy contribution typeVisual galleryRules & templateWriting patternsPaper studies

← ICCV 2025 oral writing corpus

Weiming Zhang, Dingwen Xiao, Lei Chen, Lin Wang · 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: Domain-grounded methods & physical systems (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

10 physical pages

Before references / appendix

≈ 8.05 page-equivalents

Main figures / tables

7 / 5

Attached appendix figures / tables

0 / 0

Attached appendix span

No boundary detected; not assumed zero

Before references / appendix: ≈ 80.5%References / intervening material: ≈ 19.5%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	≈ 1.06 pp.	10.6%
2. Related Works	2	≈ 0.91 pp.	9.1%
3. Methodology	3	≈ 1.97 pp.	19.7%
4. Experiments	5	≈ 2.98 pp.	29.8%
5. Conclusion and Future Work	8	≈ 0.27 pp.	2.7%
References	9	≈ 1.95 pp.	19.5%

All detected figure and table locations

Main text

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

Attached appendix

No recognized caption labels.

Argument and organization

1. 1 Introduction; 2 Related Works

   Defines class-agnostic entity segmentation and identifies over/under-segmentation in automatic SAM masks.

2. 3 Methodology

   Organizes multilevel mask generation, split-then-merge entity refinement and additional prompting of missed regions.

3. 4 Experiments; 5 Conclusion and Future Work

   Separates benchmark/backbone comparisons, module and threshold ablations, runtime and qualitative transfer.

Problem, claim and contribution

The gap is entity-level granularity and mask overlap rather than merely obtaining more masks from SAM.

Introduction and Figure 2, physical pp.1–2 · Main paper

Argument and section order

The three named modules map to three specific failure modes and are revisited in both a sequential visual example and component ablations.

Sections 3.2–3.4, pp.4–5; Figure 6 and Table 3, p.7 · Main paper

Figures, captions and visual explanation

The overview labels 32×32 versus 64×64 point grids, mask granularities and the return path for additional prompts, exposing multiple inference passes.

Figure 3, p.3 · Main paper

Tables, comparisons and evidence

The headline 30.1-point gain is the ViT-H high-resolution comparison with SAM; separate low-resolution/backbone tables avoid treating that as every setting’s improvement.

Tables 1–2, p.6; comparison text, p.7 · Main paper

Appendix and reproducibility

Dataset details, extra visualizations, hyperparameter analysis, robustness and failure cases are routed to supplementary material, not inspected.

Sections 4.1–4.4, pp.5,7–8 · Main paper

Clarity and writing conventions

The runtime discussion acknowledges 9.84 seconds per high-resolution image and future work targets speed, keeping training savings separate from inference latency.

Table 5 and Sections 4.4–5, p.8 · Main paper

What to reuse—and what not to infer

Transferable lesson: Tie each post-processing module to a visible failure mode and disclose its extra inference work alongside training-free claims.

Boundary: No additional training does not mean no pretrained supervision, no hyperparameters or fast inference. The title’s “every entity” is a task aspiration, not a completeness guarantee; qualitative open-world examples and the uninspected supplementary failure study cannot certify all unseen entities.

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

3, 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 · 10 pages