Writing study: AnchorFlow: Training-Free 3D Editing via Latent Anchor-Aligned Flows

CVPR overviewCorpus atlasBy contribution typeVisual galleryRules & templateWriting patternsPaper studies

← CVPR 2026 oral writing corpus

Zhenglin Zhou, Fan Ma, Chengzhuo Gui, Xiaobo Xia, Hehe Fan, Yi Yang, Tat-seng Chua · Official CVPR 2026 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

≈ 8.09 page-equivalents

Main figures / tables

8 / 2

Attached appendix figures / tables

0 / 0

Attached appendix span

Separate file recorded below; not in this PDF total

Before references / appendix: ≈ 73.55%References / intervening material: ≈ 26.45%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	≈ 0.82 pp.	7.45%
2. Related Work	2	≈ 0.51 pp.	4.64%
3. Method	3	≈ 2.18 pp.	19.82%
4. Experiment	5	≈ 3.44 pp.	31.27%
5. Conclusion	8	≈ 0.19 pp.	1.73%
Acknowledgments	9	≈ 0.04 pp.	0.36%
References	9	≈ 2.91 pp.	26.45%

All detected figure and table locations

Main text

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

Attached appendix

No recognized caption labels.

Separate source files—excluded from canonical-PDF totals

A separate author version may repeat main text and references. Whole-file lengths and counts are not extra appendix material and are never added to the canonical PDF totals. Read the study's source scope for which sections received substantive reading versus layout inspection.

Additional source 1

Separate PDF (or ZIP containing it): whole-file layout of 10 physical pages; 4 detected figures and 1 tables. These may include overlapping main-text material.

Located appendix only: starts on physical p.1; ≈ 9.91 page-equivalents, 4 figure labels and 1 table labels. This interval is distinct from the whole file.

Detected heading	PDF page	Estimated span
Appendix	1	≈ 0.21 pp.
A. Derivation of the Latent Anchor-based Optimization	1	≈ 1.01 pp.
B. Additional Implementation Details	2	≈ 2.28 pp.
C. Supplementary Experimental Settings	4	≈ 2.12 pp.
D. Additional Experiments	6	≈ 4.3 pp.

• Table 3 · p. 7
• Figure 9 · p. 7
• Figure 10 · p. 8
• Figure 11 · p. 9
• Figure 12 · p. 10

Argument and organization

1. Failure diagnosis to anchor formulation

   Contrast random, fixed and aligned anchors before deriving the update.

2. Sampling algorithm and categorized evidence

   Give executable-style sampling steps, evaluate five edit categories, then expose parameter effects and reconstruction limits.

Problem, claim and contribution

The introduction contrasts under-editing from inconsistent anchors with over-editing from fixed anchors; the intervention is motivated as a balance between two visible failure modes.

Figure 2 and Section 3.2, PDF pp. 2–3 · Main paper · Additional source 2

Argument and section order

Flow-matching preliminaries and the global-anchor definition precede Algorithm 1, separating the mathematical idea from its sampling implementation.

Sections 3.1–3.3 and Algorithm 1, PDF pp. 3–5 · Main paper · Additional source 2

Figures, captions and visual explanation

The qualitative grid keeps condition pairs and the source object beside every method, with each row's instruction printed immediately below it; readers can inspect both the changed part and preserved identity.

Figure 4, PDF p. 6; visual inspection · Main paper · Additional source 2

Tables, comparisons and evidence

The results distinguish five edit categories, runtime, averaging and parameter sensitivity; the paper explicitly notes that removal performance is moderate rather than hiding that subtype in the overall mean.

Table 1, Figures 5–7 and Table 2, PDF pp. 7–8 · Main paper · Additional source 2

Appendix and reproducibility

The supplement provides a navigable A–D roadmap for derivation, dataset creation, metrics/baseline setup and additional experiments, including the prompts used in constructing evaluation conditions.

Supplement opening roadmap and Appendices A–D, PDF pp. 1–10 · Main paper · Additional source 2

Clarity and writing conventions

A dedicated limitation subsection identifies the upstream VAE reconstruction ceiling, separating an inherited representation failure from the editing update itself.

Section 4.4 and Figure 8, PDF p. 8 · Main paper · Additional source 2

What to reuse—and what not to infer

Transferable lesson: For an inference-only method, link a concrete failure diagnostic to the update equation, provide pseudocode, and report quality, cost and sensitivity as separate tests.

Boundary: The identity-preservation score uses a generated target image as reference under an explicit assumption; it is not a direct geometric ground-truth guarantee. Keep the moderate object-removal result visible.

Apply this lesson using the argument and evidence worksheet.

Inspection record and sources

Checked

2026-09-26

Reading status

Writing-focused manuscript scan + selected visuals inspected

Main text

Manuscript-wide writing scan of PDF pp. 1–11: framing, argument sequence, section organization, figure/table captions, evaluation interpretation and limitations; reference sections identified. This is a writing-focused scan, not an exhaustive technical proof or reproduction audit. Main PDF pages 3 and 6 were rendered and visually inspected.

Appendix / supplement

Writing-focused scan of the separate supplementary PDF, pp. 1–10, including its organization, captions, protocol/evidence detail and limitation material where present; proofs and code listings were inspected for exposition, not independently verified or executed. Supplement opening roadmap and Appendices A–D, PDF pp. 1–10. Linked video/audio demonstrations and non-PDF archive contents were not assessed.

Visually inspected PDF pages

3, 6

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 CVPR 2026 CVF open-access accepted manuscript; retrieved 2026-09-26 · 11 pages
• Official CVPR 2026 CVF supplementary PDF; retrieved 2026-09-26 · 10 pages