Skip to content
Computational Pathology Reading Roadmap

Writing study: GaussianGPT: Towards Autoregressive 3D Gaussian Scene Generation

← ECCV 2026 oral writing corpus

Nicolas von Lützow, Barbara Roessle, Katharina Schmid, Matthias Niessner · Official ECCV 2026 oral record

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
21 physical pages
Before references / appendix
≈ 15.3 page-equivalents
Main figures / tables
6 / 4
Attached appendix figures / tables
0 / 0
Attached appendix span
No boundary detected; not assumed zero
Before references / appendix: ≈ 72.86%References / intervening material: ≈ 27.14%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.

Source-bound correction: All top-level main headings and the reference boundary checked against source PDF text; source-bound positions use physical single-column y/page-height, not a two-column reading-order surrogate. Running page-number/author titles, affiliations, numbered contribution bullets, footnotes and bibliography continuations are excluded. This linked manuscript contains no appended supplementary section; separate supplement links were not scanned.

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 headingStarts on PDF pageEstimated spanShare of entire PDF
1 Introduction2≈ 1 pp.4.76%
2 Related Works3≈ 1.27 pp.6.05%
3 Methodology4≈ 4.09 pp.19.48%
4 Experiments8≈ 6.91 pp.32.9%
5 Conclusion15≈ 0.73 pp.3.48%
Acknowledgements16≈ 0.15 pp.0.71%
References16≈ 5.7 pp.27.14%
All detected figure and table locations

Main text

Attached appendix

    No recognized caption labels.

    Argument and organization

    1. 1 Introduction; 2 Related Works

      Presents autoregression as an alternative generative interface with completion and outpainting flexibility, not simply a claim that diffusion is universally inferior.

    2. 3 Methodology: scene compression and autoregressive latent grids

      Explains continuous-scene compression before serialization and prediction; evaluation separates objects, matched scene chunks, completion and extended outpainting.

    3. 4 Experiments; 5 Conclusion

      The quantitative scene table reports different sample counts for generation and completion, while token-order and vocabulary/RoPE ablations are evaluated separately.

    Problem, claim and contribution

    Presents autoregression as an alternative generative interface with completion and outpainting flexibility, not simply a claim that diffusion is universally inferior.

    Abstract and Section 1, pp.1–3 · Main paper

    Argument and section order

    Explains continuous-scene compression before serialization and prediction; evaluation separates objects, matched scene chunks, completion and extended outpainting.

    Sections 3–4, pp.4–15 · Main paper

    Figures, captions and visual explanation

    A two-tier overview shows compression above interleaved position/feature prediction; the outpainting panel displays extended scenes with a physical size in the caption.

    Figure 2, p.5; Figure 5, p.13 · Main paper

    Tables, comparisons and evidence

    The quantitative scene table reports different sample counts for generation and completion, while token-order and vocabulary/RoPE ablations are evaluated separately.

    Table 2, p.12; Tables 3–4, p.15 · Main paper

    Appendix and reproducibility

    Named appendix pointers cover autoencoder hyperparameters, matched chunk baselines and degradation away from initial context. The separate appendix was not retrieved.

    Sections 4.2–4.5, pp.9,11,13 · Main paper

    Clarity and writing conventions

    The text discloses that long-range generation quality gradually degrades and separates full-room qualitative comparisons from matched chunk-level quantitative tests.

    Sections 4.4–4.5, pp.11–13 · Main paper

    What to reuse—and what not to infer

    Transferable lesson: For autoregressive representations, show serialization explicitly and evaluate unconditional generation, conditioning and extrapolated context under their own protocols.

    Boundary: A larger outpainted scene is not evidence of unchanged quality outside the training horizon, and full-room versus chunk comparisons require explicit qualification.

    Apply this lesson using the argument and evidence worksheet.

    Inspection record and sources

    Checked
    2026-10-02
    Reading status
    Writing-focused manuscript scan + selected visuals inspected
    Main text
    Writing-oriented scan across the manuscript main argument, physical PDF pp.1–15: framing and section roadmaps, method/argument explanations, figure and table captions, comparison protocols, conclusions and limitations. Selected full-page rasters were inspected. This is not exhaustive line-by-line technical reading, proof verification or experimental reproduction.
    Appendix / supplement
    Main-manuscript pointers to appendix/supplement and the reference boundary were checked. Separate supplementary PDFs were not retrieved or inspected; their contents are not included in inspected-page counts.
    Visually inspected PDF pages
    5, 13

    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.