Writing study: Back on Track: Bundle Adjustment for Dynamic Scene Reconstruction

ICCV overviewCorpus atlasBy contribution typeVisual galleryRules & templateWriting patternsPaper studies

← ICCV 2025 oral writing corpus

Weirong Chen, Ganlin Zhang, Felix Wimbauer, Rui Wang, Nikita Araslanov, Andrea Vedaldi, Daniel Cremers · 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

10 physical pages

Before references / appendix

≈ 8.11 page-equivalents

Main figures / tables

7 / 4

Attached appendix figures / tables

0 / 0

Attached appendix span

No boundary detected; not assumed zero

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	≈ 1.3 pp.	13%
2. Related Work	2	≈ 0.73 pp.	7.3%
3. BA-Track	3	≈ 2.52 pp.	25.2%
4. Experiments	5	≈ 2.92 pp.	29.2%
5. Conclusion	8	≈ 0.28 pp.	2.8%
References	9	≈ 1.89 pp.	18.9%

All detected figure and table locations

Main text

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

Attached appendix

No recognized caption labels.

Argument and organization

1. 1 Introduction; 2 Related Work

   Explains why discarding dynamic points loses structure and why independent motion estimates misalign geometry.

2. 3 BA-Track

   Orders a learned motion-decoupling tracker, bundle adjustment and global dense-depth refinement as three stages.

3. 4 Experiments; 5 Conclusion

   Separates pose, depth, motion-decoupling and dynamic reconstruction evidence.

Problem, claim and contribution

Preserves the familiar bundle-adjustment contract by estimating the camera-induced motion of dynamic points, rather than claiming standard epipolar geometry directly handles arbitrary moving objects.

Sections 1 and 3.1, physical pp.1–4 · Main paper

Argument and section order

The method follows sparse tracking to pose/geometry optimization to dense depth refinement; each stage has its own variables and loss.

Sections 3.1–3.3, pp.3–5 · Main paper

Figures, captions and visual explanation

A three-stage overview labels online sparse dynamic SLAM separately from final dense reconstruction and maps predicted motion, visibility and dynamic labels into bundle adjustment.

Figure 3, p.3 · Main paper

Tables, comparisons and evidence

Pose and depth tables mark subsequence evaluation caused by memory limits; the depth table also identifies test-time finetuning and image-pair depth privileges.

Tables 1–2, pp.6–7 · Main paper

Appendix and reproducibility

The implementation section provides key tracker/BA settings in the main text and explicitly routes further details to supplementary material; only the main pointers were inspected.

Section 4.1, p.6 · Main paper

Clarity and writing conventions

The results distinguish strongest ATE from competitive RTE/RRE, avoiding an all-metrics-win summary. Pose evaluation is explicitly after Sim(3) alignment.

Section 4.2, pp.6–7 · Main paper

What to reuse—and what not to infer

Transferable lesson: Use stage-specific contracts and tables to keep camera motion, object motion and depth quality from collapsing into one reconstruction claim.

Boundary: Aligned trajectory errors do not measure absolute metric-scale recovery. Subsequence results have a different evaluation scope; alternate-depth qualitative examples alone do not establish independence from depth-prior quality.

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, 7

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