Writing study: 3D-LATTE: Latent Space 3D Editing from Textual Instructions

CVPR overviewCorpus atlasBy contribution typeVisual galleryRules & templateWriting patternsPaper studies

← CVPR 2026 oral writing corpus

Maria Parelli, Michael Oechsle, Michael Niemeyer, Federico Tombari, Andreas Geiger · 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

10 physical pages

Before references / appendix

≈ 8.11 page-equivalents

Main figures / tables

9 / 2

Attached appendix figures / tables

0 / 0

Attached appendix span

Separate file recorded below; not in this PDF total

Before references / appendix: ≈ 81.1%References / intervening material: ≈ 18.9%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	≈ 0.8 pp.	8%
2. Related Works	2	≈ 0.59 pp.	5.9%
3. Method	3	≈ 3.18 pp.	31.8%
4. Experiments	6	≈ 2.43 pp.	24.3%
5. Conclusion	8	≈ 0.25 pp.	2.5%
References	9	≈ 1.89 pp.	18.9%

All detected figure and table locations

Main text

• Figure 1 · p. 1
• Figure 2 · p. 4
• Figure 3 · p. 4
• Figure 4 · p. 5
• Table 1 · p. 6
• Figure 5 · p. 7
• Table 2 · p. 7
• Figure 6 · p. 8
• Figure 7 · p. 8
• Figure 8 · p. 8
• Figure 9 · 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 9 physical pages; 6 detected figures and 4 tables. These may include overlapping main-text material.

Located appendix only: starts on physical p.1; ≈ 8.91 page-equivalents, 6 figure labels and 4 table labels. This interval is distinct from the whole file.

Detected heading	PDF page	Estimated span
Supplementary Material	1	≈ 0.22 pp.
1. Additional Results	1	≈ 0.07 pp.
2. Application for geometry control	1	≈ 0.31 pp.
3. Experimental Setup	1	≈ 5.73 pp.
4. Mask Generation	7	≈ 0.74 pp.
5. Runtime cost	8	≈ 0.21 pp.
6. Further ablations on 3D attention injection	8	≈ 0.48 pp.
7. Generalization across 3D representations	8	≈ 0.21 pp.
8. Limitations	9	≈ 0.12 pp.
9. Broader Impact	9	≈ 0.07 pp.
References	9	≈ 0.75 pp.

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

Argument and organization

1. Related work and representation

   Contrast 2D-prior and multi-view approaches before defining Gaussian latent editing.

2. Method to evaluation

   Introduce inversion, attention, masks and refinement before testing edit fidelity and preservation.

Problem, claim and contribution

The opening distinguishes the failure of 2D editing signals from the desired globally consistent geometric change, making the choice of a native 3D latent space the argumentative pivot.

Abstract and Sections 1–2, PDF pp. 1–3 · Main paper · Additional source 2

Argument and section order

Section 3 starts with a roadmap to each component and separates the inherited DiffSplat representation from the proposed editing operations.

Section 3 and Sections 3.1–3.7, PDF pp. 3–6 · Main paper · Additional source 2

Figures, captions and visual explanation

The compact result table on page 6 explicitly gives opposite metric directions for editing alignment and unchanged-region difference; rank colors complement bold numbers rather than replacing metric labels.

Table 1, PDF p. 6; visual inspection · Main paper · Additional source 2

Tables, comparisons and evidence

CLIP metrics, pairwise GPTEval3D preferences and a human study supply different evidence for instruction following, view consistency and visual quality; these should not be collapsed into one accuracy claim.

Tables 1–2 and Figure 7, PDF pp. 6–8 · Main paper · Additional source 2

Appendix and reproducibility

The supplement separates additional qualitative comparisons, geometry-control use cases and the metric/implementation protocol, making the extra images interpretable as evidence rather than decoration.

Supplement Sections 1–3, PDF pp. 1–6 · Main paper · Additional source 2

Clarity and writing conventions

Named component subsections and a method roadmap help readers navigate a pipeline that would otherwise appear as an unstructured collection of improvements.

Section 3, PDF pp. 3–6 · Main paper · Additional source 2

What to reuse—and what not to infer

Transferable lesson: For editing papers, present separate evidence for the intended change and preservation of untouched content, and show exactly which stage enforces each.

Boundary: Do not turn the best average alignment into universal superiority: Table 1's unchanged-region difference is lower for GaussCTRL than for 3D-LATTE.

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. 1–10: 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. 1–9, 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 Sections 1–3, PDF pp. 1–6. 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 · 10 pages
• Official CVPR 2026 CVF supplementary PDF; retrieved 2026-09-26 · 9 pages