Writing study: EVER: Exact Volumetric Ellipsoid Rendering for Real-time View Synthesis

ICCV overviewCorpus atlasBy contribution typeVisual galleryRules & templateWriting patternsPaper studies

← ICCV 2025 oral writing corpus

Alexander Mai, Peter Hedman, George Kopanas, Dor Verbin, David Futschik, Qiangeng Xu, Falko Kuester, Jonathan T. Barron, Yinda Zhang · 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: Efficiency, inference & 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 / 3

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	≈ 0.45 pp.	4.5%
2. Related Work	2	≈ 0.49 pp.	4.9%
3. Motivation	2	≈ 0.49 pp.	4.9%
4. Method	3	≈ 1.57 pp.	15.7%
5. Implementation	4	≈ 0.87 pp.	8.7%
6. Results	5	≈ 3.02 pp.	30.2%
7. Conclusion	8	≈ 0.31 pp.	3.1%
References	9	≈ 1.95 pp.	19.5%

All detected figure and table locations

Main text

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

Attached appendix

No recognized caption labels.

Argument and organization

1. 1 Introduction; 2 Related Work; 3 Motivation

   Isolates popping as a consequence of approximate overlap and sorting, rather than treating all rendering artifacts as one problem.

2. 4 Method; 5 Implementation

   Explains constant-density ellipsoids and exact interval blending before ray sorting and density-control changes.

3. 6 Results; 7 Conclusion

   Pairs reconstruction metrics and costs with a popping-specific test and ablations of blending, splitting and regularization.

Problem, claim and contribution

The exactness claim concerns volumetric integration for constant-density primitives, with real-time performance reported at an explicit resolution and GPU.

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

Argument and section order

A flatland failure example precedes the interval algorithm, and implementation changes are separated from the rendering derivation.

Sections 3–5, pp.2–5 · Main paper

Figures, captions and visual explanation

A two-primitive cartoon is followed by sorted entry/exit events and piecewise density intervals; an epipolar-image diagnostic makes discontinuities visible under camera movement.

Figures 2–3, p.3 · Main paper

Tables, comparisons and evidence

The main comparison separates image metrics from GPU-hours, disk model size and FPS; its caption flags inherited timings using different GPUs rather than implying all measurements are matched.

Table 1, p.6 · Main paper

Appendix and reproducibility

Density-control detail is linked to Appendix B and temporal popping to a supplemental video. Those pointers were checked, but neither separate content source was inspected.

Section 5, p.4; Section 6.2, p.6 · Main paper

Clarity and writing conventions

The caption defines Mem. as finished-model disk size, preventing readers from assuming GPU peak memory. The text distinguishes test-resolution FPS from the 720p headline.

Table 1, p.6; Section 6.3, p.7 · Main paper

What to reuse—and what not to infer

Transferable lesson: Use the smallest worked construction to explain an exact algorithm, then measure the failure mode separately from average image quality.

Boundary: Exact integration is conditional on the chosen primitive model, not proof that reconstructed geometry or illumination is physically correct. Cross-paper GPU-hours/FPS rows with different hardware cannot establish a fully matched speedup.

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, 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 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