Writing study: RayZer: A Self-supervised Large View Synthesis Model

ICCV overviewCorpus atlasBy contribution typeVisual galleryRules & templateWriting patternsPaper studies

← ICCV 2025 oral writing corpus

Hanwen Jiang, Hao Tan, Peng Wang, Haian Jin, Yue Zhao, Sai Bi, Kai Zhang, Fujun Luan, Kalyan Sunkavalli, Qixing Huang, Georgios Pavlakos · 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: Training, objectives & optimization (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

12 physical pages

Before references / appendix

≈ 8.08 page-equivalents

Main figures / tables

6 / 7

Attached appendix figures / tables

0 / 0

Attached appendix span

No boundary detected; not assumed zero

Before references / appendix: ≈ 67.33%References / intervening material: ≈ 32.67%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	≈ 0.92 pp.	7.67%
2. Related Work	2	≈ 0.88 pp.	7.33%
3. Preliminaries	3	≈ 0.32 pp.	2.67%
4. RayZer	3	≈ 2.19 pp.	18.25%
5. Experiments	5	≈ 2.68 pp.	22.33%
6. Conclusion	8	≈ 0.65 pp.	5.42%
References	9	≈ 3.92 pp.	32.67%

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

Attached appendix

No recognized caption labels.

Argument and organization

1. 1 Introduction; 2 Related Work; 3 Preliminaries

   Introduces self-supervised camera/scene disentanglement and the latent-set rendering representation.

2. 4 RayZer

   Separates the training information split from the concrete camera-first transformer, ray conditioning and rendering decoder.

3. 5 Experiments; conclusion

   Distinguishes supervision regimes, ordered/unordered inputs, learned-pose geometry and architecture ablations.

Problem, claim and contribution

Defines zero 3D supervision specifically as no pose or geometry labels, while retaining multiview images and a ray-based geometric prior.

Abstract and Sections 1,4.1, physical pp.1–3 · Main paper

Argument and section order

An abstract two-set self-supervision diagram precedes the operational camera-to-rays-to-scene pipeline, reducing the burden of learning the objective and architecture at once.

Figure 2, p.2; Sections 4.1–4.2, pp.3–5 · Main paper

Figures, captions and visual explanation

The three-panel overview labels input and target view sets, a canonical camera and predicted ray maps, making the target-image information path inspectable.

Figure 3, p.4 · Main paper

Tables, comparisons and evidence

The pose-awareness table explicitly changes protocol from rendering at estimated target poses to interpolating input poses using coefficients from ground-truth poses, and includes nearest-view copying as a control.

Table 5 and Section 5.3, p.8 · Main paper

Appendix and reproducibility

Architecture/training details, dataset settings and pose interpolation details are deferred to the appendix; only these main-manuscript pointers were inspected.

Sections 5.1 and 5.3, pp.5,7 · Main paper

Clarity and writing conventions

The pose analysis states that learned SE(3) values need not match real-world pose space; image rendering quality is therefore not silently equated with metric pose accuracy.

Section 5.3, p.7 · Main paper

What to reuse—and what not to infer

Transferable lesson: Draw every information path in a self-supervised objective, then separate representation usefulness from external geometric accuracy.

Boundary: Target images are used for pose estimation in the main rendering protocol. Oracle-labelled baselines and the interpolation test have distinct information access; no-pose-label training is not no-geometric-prior learning.

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 · 12 pages