ECCV 2026 — corpus atlas
All 28 oral main manuscripts in the official program snapshot were scanned. Separate supplementary PDFs were not retrieved; main-text pointers are not a substitute for reading them.
A figure count means a distinct caption label, not a panel, image or graph count. A six-panel figure counts once. There is no figure quota or implied acceptance threshold. Use the type guides to understand each artifact’s purpose and the gallery to see selected styles directly.
What the measured manuscripts look like
Observed extraction results, not a target length or figure quota. The middle 50% is the interval from the 25th to the 75th percentile. Missing boundaries are excluded from that span’s denominator, not replaced by zero. Caption counts are extraction-assisted and can miss nonstandard or image-only labels; inspect each paper’s notes.
| Measure | Median | Middle 50% | Full range | Denominator |
|---|---|---|---|---|
| Inspected PDF lengthphysical pages | 19 | 18–19.25 | 15–22 | 28 papers |
| Before references / appendixestimated page-equivalents | 15.15 | 14.59–15.15 | 12.51–15.3 | 28 papers |
| Attached appendix spanestimated page-equivalents | Not resolved | Not resolved–Not resolved | Not resolved–Not resolved | 0 papers |
| Main figurescaption labels | 7 | 5–9 | 3–13 | 28 papers |
| Main tablescaption labels | 4 | 2–6 | 0–9 | 28 papers |
| Appendix figurescaption labels | 0 | 0–0 | 0–0 | 28 papers |
| Appendix tablescaption labels | 0 | 0–0 | 0–0 | 28 papers |
| Separate source-file lengthwhole-file physical pages, possibly including repeated main text; not appendix-only length | Not resolved | Not resolved–Not resolved | Not resolved–Not resolved | 0 papers |
| Separate source-file figureswhole-file caption labels; may include overlapping main-text displays | Not resolved | Not resolved–Not resolved | Not resolved–Not resolved | 0 papers |
| Separate source-file tableswhole-file caption labels; may include overlapping main-text displays | Not resolved | Not resolved–Not resolved | Not resolved–Not resolved | 0 papers |
Compare like contributions
Each measured paper has one primary editorial type so the totals sum to 28. Secondary contributions were not separately coded; empty secondary lists do not imply the absence of mixed contributions. These are classifications of the annotated argument, not official ECCV tracks. The table uses the same extraction method and shows medians; the category guides below explain what these artifacts should accomplish.
| Primary contribution | Papers | Main figures / tables | Body span ≈ pp. | Appendix span ≈ pp. |
|---|---|---|---|---|
| Architecture & representation | 10 | 6.5 / 4 | 15.15 | Not resolved0/10 attached boundaries; separate PDF median Not resolved pp. (0 papers) |
| Training, objectives & optimization | 4 | 5.5 / 5.5 | 15.15 | Not resolved0/4 attached boundaries; separate PDF median Not resolved pp. (0 papers) |
| Theory & formal analysis | 2 | 11 / 0.5 | 14.97 | Not resolved0/2 attached boundaries; separate PDF median Not resolved pp. (0 papers) |
| Empirical analysis, diagnosis & auditing | 4 | 7 / 2.5 | 14.63 | Not resolved0/4 attached boundaries; separate PDF median Not resolved pp. (0 papers) |
| Datasets, benchmarks & measurement | 2 | 5.5 / 4.5 | 13.51 | Not resolved0/2 attached boundaries; separate PDF median Not resolved pp. (0 papers) |
| Efficiency, inference & systems | 5 | 5 / 4 | 15.15 | Not resolved0/5 attached boundaries; separate PDF median Not resolved pp. (0 papers) |
| Domain-grounded methods & physical systems | 1 | 9 / 0 | 15.15 | Not resolved0/1 attached boundaries; separate PDF median Not resolved pp. (0 papers) |
Search all 28 measurement records
Open a paper’s ledger for every detected top-level section, its estimated length and share of the PDF, and figure/table locations. Its existing study explains what those visuals and sections do.
Showing 28 of 28 records.
| Paper / category | PDF pages | Main F / T | Appendix F / T | Body / appendix span ≈ pp. | Notes |
|---|---|---|---|---|---|
| Provable and Robust Wavefront Sensing via Self-Reference InterferometryTheory & formal analysis | 19 | 9 / 1 | 0 / 0 | 15.15 / Unresolved | No attached appendix boundary detectedSource correction retained |
| Broadband Wide Field of View Imaging with Computational MirrorsDomain-grounded methods & physical systems | 18 | 9 / 0 | 0 / 0 | 15.15 / Unresolved | No attached appendix boundary detectedSource correction retained |
| Poppy: Polarization-Based Plug-and-Play Guidance for Enhancing Surface Normal EstimationTraining, objectives & optimization | 19 | 10 / 1 | 0 / 0 | 15.15 / Unresolved | No attached appendix boundary detectedSource correction retained |
| A second-order theory of texture for depth from focusTheory & formal analysis | 20 | 13 / 0 | 0 / 0 | 14.79 / Unresolved | No attached appendix boundary detectedSource correction retained |
| PRISM-VO: Scale-Aware Visual Odometry Using Photometric Plenoptic Bundle AdjustmentArchitecture & representation | 18 | 12 / 2 | 0 / 0 | 15.15 / Unresolved | No attached appendix boundary detectedSource correction retained |
| Cross-View Yaw Estimation in Location Uncertainty with Line-Aligning Yaw ScoringArchitecture & representation | 18 | 5 / 8 | 0 / 0 | 15.15 / Unresolved | No attached appendix boundary detectedSource correction retained |
| Delving into Latent Spectral Biasing of Video VAEs for Superior DiffusabilityTraining, objectives & optimization | 18 | 6 / 6 | 0 / 0 | 15.15 / Unresolved | No attached appendix boundary detectedSource correction retained |
| OmniScript: Towards Audio-Visual Script Generation for Long-Form Cinematic VideoDatasets, benchmarks & measurement | 17 | 4 / 5 | 0 / 0 | 14.51 / Unresolved | No attached appendix boundary detectedSource correction retained |
| OmniForcing: Unleashing Real-time Joint Audio-Visual GenerationEfficiency, inference & systems | 19 | 4 / 4 | 0 / 0 | 15.15 / Unresolved | No attached appendix boundary detectedSource correction retained |
| Keep It Simple: Multi-Key Episodic Memory Retrieval for Ultra-Long Video UnderstandingEfficiency, inference & systems | 18 | 3 / 7 | 0 / 0 | 15.15 / Unresolved | No attached appendix boundary detectedSource correction retained |
| FEEL (Force-Enhanced Egocentric Learning): A Dataset for Physical Action UnderstandingDatasets, benchmarks & measurement | 15 | 7 / 4 | 0 / 0 | 12.51 / Unresolved | No attached appendix boundary detectedSource correction retained |
| Steerable Vision TransformersArchitecture & representation | 17 | 10 / 4 | 0 / 0 | 14.32 / Unresolved | No attached appendix boundary detectedSource correction retained |
| Understanding Geometric Representations in Self-Supervised Vision Transformers via Subspace InterventionEmpirical analysis, diagnosis & auditing | 18 | 8 / 2 | 0 / 0 | 14.59 / Unresolved | No attached appendix boundary detectedSource correction retained |
| World Knowledge in the Weights: Reading Concept Circuits of Vision TransformersEmpirical analysis, diagnosis & auditing | 19 | 7 / 2 | 0 / 0 | 14.35 / Unresolved | No attached appendix boundary detectedSource correction retained |
| What CLIP Knows but Cannot Say: Recovering Negation from Frozen Intermediate FeaturesEmpirical analysis, diagnosis & auditing | 19 | 7 / 6 | 0 / 0 | 15.15 / Unresolved | No attached appendix boundary detectedSource correction retained |
| Chain-of-Visual-Thought: Teaching VLMs to See and Think Better with Continuous Visual TokensArchitecture & representation | 20 | 9 / 6 | 0 / 0 | 14.37 / Unresolved | No attached appendix boundary detectedSource correction retained |
| Make Geometry Matter for Spatial ReasoningTraining, objectives & optimization | 19 | 3 / 7 | 0 / 0 | 14.79 / Unresolved | No attached appendix boundary detectedSource correction retained |
| Heat Kernel Textures -- the Geodesic Gaussians That Do Not SplatArchitecture & representation | 19 | 4 / 2 | 0 / 0 | 14.69 / Unresolved | No attached appendix boundary detectedSource correction retained |
| NURBS Splatting: A Unified Differentiable Rendering Framework for Vector GraphicsEfficiency, inference & systems | 18 | 7 / 2 | 0 / 0 | 14.15 / Unresolved | No attached appendix boundary detectedSource correction retained |
| TriFlow: Generating Artist-Like 3D Mesh Topology via Nearest-Vertex Vector FieldsArchitecture & representation | 19 | 7 / 3 | 0 / 0 | 14.59 / Unresolved | No attached appendix boundary detectedSource correction retained |
| Repurposing Geometric Foundation Models for Multi-view DiffusionArchitecture & representation | 19 | 7 / 9 | 0 / 0 | 15.15 / Unresolved | No attached appendix boundary detectedSource correction retained |
| MIMFlow: Integrating Masked Image Modeling with Normalizing Flows for End-to-End Image GenerationTraining, objectives & optimization | 18 | 5 / 5 | 0 / 0 | 15.15 / Unresolved | No attached appendix boundary detectedSource correction retained |
| Register Any Point: Scaling 3D Point Cloud Registration by Flow MatchingArchitecture & representation | 22 | 6 / 3 | 0 / 0 | 15.15 / Unresolved | No attached appendix boundary detectedSource correction retained |
| Geometric Context Transformer for Streaming 3D ReconstructionEfficiency, inference & systems | 21 | 5 / 7 | 0 / 0 | 15.15 / Unresolved | No attached appendix boundary detectedSource correction retained |
| LSRM: High-Fidelity Object-Centric Reconstruction via Scaled Context WindowsEfficiency, inference & systems | 21 | 9 / 4 | 0 / 0 | 15.15 / Unresolved | No attached appendix boundary detectedSource correction retained |
| GaussianGPT: Towards Autoregressive 3D Gaussian Scene GenerationArchitecture & representation | 21 | 6 / 4 | 0 / 0 | 15.3 / Unresolved | No attached appendix boundary detectedSource correction retained |
| Volume Transformer: Revisiting Vanilla Transformers for 3D Scene UnderstandingArchitecture & representation | 20 | 4 / 4 | 0 / 0 | 15.15 / Unresolved | No attached appendix boundary detectedSource correction retained |
| Seeing Through the Weights: Privacy Leakage in Scene Coordinate RegressionEmpirical analysis, diagnosis & auditing | 18 | 7 / 3 | 0 / 0 | 14.67 / Unresolved | No attached appendix boundary detectedSource correction retained |
No matching papers. Reset or use a shorter query.
Measurement method
Section titled “Measurement method”ECCV’s single-column LNCS pages are measured in physical-page coordinates. Running headers, affiliation lines, numbered contribution bullets and bibliography entries were excluded through a source-linked heading audit of all 28 papers. Caption inventories were independently cross-checked for seven papers, not every label in the corpus.
Each top-level section extends from its detected heading to the next detected heading. Its share is relative to the entire inspected PDF, not the conference content-page allowance. These are approximate layout envelopes, including whitespace and floats—not exact text lengths, word counts, occupied-ink area or official page-limit accounting. The title, abstract and teaser may precede the first numbered section.
Missing appendix boundaries remain unresolved rather than zero. Separate files have separate denominators. Where a checklist boundary is resolved, it ends the preceding appendix span; checklist pages are not silently treated as technical appendix content. Source revisions, extraction notes and URL-linked corrections remain in each paper ledger.
Not every caption was hand-counted. Counts are extraction-assisted; nonstandard or image-only captions, repeated labels, contents pages and atypical text layers can need corrections. The source page remains authoritative. Category assignments are editorial labels from the writing studies, not official conference tracks; secondary contributions were not separately coded.
Selected examples do not establish visual-style prevalence. An accepted-only collection cannot identify which writing choices caused acceptance.
Download structural metadata · Inspect source status and studies.