Writing study: Learning Latent Concepts for Detecting Out-of-Distribution Objects

CVPR overviewCorpus atlasBy contribution typeVisual galleryRules & templateWriting patternsPaper studies

← CVPR 2026 oral writing corpus

Ting Peng, Junhao Dong, Yew-Soon Ong · 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

11 physical pages

Before references / appendix

≈ 8.13 page-equivalents

Main figures / tables

5 / 6

Attached appendix figures / tables

0 / 0

Attached appendix span

Separate file recorded below; not in this PDF total

Before references / appendix: ≈ 73.91%References / intervening material: ≈ 26.09%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	≈ 1.26 pp.	11.45%
2. Related Work	2	≈ 0.77 pp.	7%
3. Proposed Method	3	≈ 2.46 pp.	22.36%
4. Experiments	5	≈ 3.02 pp.	27.45%
5. Conclusion	8	≈ 0.25 pp.	2.27%
References	9	≈ 2.87 pp.	26.09%

All detected figure and table locations

Main text

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

Located appendix only: starts on physical p.1; ≈ 2.94 page-equivalents, 4 figure labels and 3 table labels. This interval is distinct from the whole file.

Detected heading	PDF page	Estimated span
Supplementary Material	1	≈ 0.02 pp.
6. OOD-OD Results on Faster R-CNN	1	≈ 0.27 pp.
7. OOD Image Classification	1	≈ 0.45 pp.
8. Additional Empirical Analysis	1	≈ 2.2 pp.

• Table 7 · p. 1
• Table 8 · p. 2
• Figure 6 · p. 2
• Figure 7 · p. 2
• Figure 8 · p. 3
• Figure 9 · p. 3
• Table 9 · p. 3

Argument and organization

1. Instance-level limitation

   Motivate global latent concepts as a complement to detector-local decisions.

2. Train concepts, bind at inference

   Separate unsupervised concept discovery from inference-time binding and scoring, then evaluate multiple OOD settings.

Problem, claim and contribution

The opening distinguishes object-level OOD detection from image classification and motivates scene context rather than only a stronger classifier score.

Introduction and Figure 1 · Main paper · Additional source 2

Argument and section order

Training-time concept discovery and inference-time binder/score are explicit subsections, so unchanged-detector claims can be inspected operationally.

Sections 3.2–3.4 · Main paper · Additional source 2

Figures, captions and visual explanation

The slot visualization keeps original image, baseline activation, new activation and segmentation assignments in consistent columns across four scenes.

Figure 3, PDF p. 7 · Main paper · Additional source 2

Tables, comparisons and evidence

Detection, open-world detection and image-level OOD results have separate tables; the classification table reports both AUROC and FPR95 with direction arrows.

Tables 1–4, PDF pp. 6–7 · Main paper · Additional source 2

Appendix and reproducibility

The supplement checks a second detector family and provides task-specific configurations and OOD-score sensitivity instead of treating one detector result as universal transfer.

Supplement Sections 6–8, pp. 1–3 · Main paper · Additional source 2

Clarity and writing conventions

Coarse representation, relational regularization and slot refinement appear as descriptive run-in headings before the final objective.

Section 3.2, PDF p. 4 · Main paper · Additional source 2

What to reuse—and what not to infer

Transferable lesson: When transferring a representation mechanism across tasks, separate each task's protocol and metrics, while using matched visual columns to explain what the mechanism changes.

Boundary: More coherent slot maps are illustrative, not proof of causal object concepts; object detection and whole-image OOD results should not be averaged into one success claim.

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 main PDF pp. 1–8: framing, section structure, method exposition, figure/table captions, evaluation organization and conclusions; not an exhaustive technical proof audit or reproduction.

Appendix / supplement

Separate supplement pp. 1–3 scanned: Faster R-CNN transfer, classification datasets/setup, score sensitivity, configuration table and additional detections/slot maps.

Visually inspected PDF pages

4, 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 CVPR 2026 CVF proceedings PDF; retrieved 2026-09-26 · 11 pages
• Official CVPR 2026 separate supplementary PDF; retrieved 2026-09-26 · 3 pages