CVPR 2026 — corpus atlas
This atlas covers all 141 oral papers in the dated official inventory. It describes the accepted public PDFs, not necessarily the anonymous versions reviewed. Figures and tables are counted by distinct caption label: a six-panel figure is one figure, not six graphs.
Use the distributions to understand space allocation and artifact variety—not to choose a figure quota or predict acceptance. The contribution-type guides explain what each kind of paper needs to establish; the visual gallery shows useful forms directly.
141 / 141inspected PDFs / oral inventory records
1,522physical pages in those PDFs
7editorial contribution types
130separate files, excluded from totals
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 | 11 | 10–11 | 9–17 | 141 papers |
| Before references / appendixestimated page-equivalents | 8.09 | 8.05–8.13 | 7.76–8.34 | 141 papers |
| Attached appendix spanestimated page-equivalents | 2.62 | 1.91–3.11 | 1.07–8.8 | 5 papers |
| Main figurescaption labels | 6 | 5–8 | 2–16 | 141 papers |
| Main tablescaption labels | 4 | 3–5 | 0–10 | 141 papers |
| Appendix figurescaption labels | 0 | 0–0 | 0–1 | 141 papers |
| Appendix tablescaption labels | 0 | 0–0 | 0–0 | 141 papers |
| Separate source-file lengthwhole-file physical pages, possibly including repeated main text; not appendix-only length | 7 | 4.25–10 | 1–62 | 130 papers |
| Separate source-file figureswhole-file caption labels; may include overlapping main-text displays | 5 | 3–8 | 0–29 | 130 papers |
| Separate source-file tableswhole-file caption labels; may include overlapping main-text displays | 3 | 2–6 | 0–27 | 130 papers |
Compare like contributions
Each measured paper has one primary editorial type so the totals sum to 141. 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 CVPR 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 | 49 | 7 / 4 | 8.09 | 1.911/49 attached boundaries; separate PDF median 6 pp. (46 papers) |
| Training, objectives & optimization | 24 | 6 / 4 | 8.1 | 1.071/24 attached boundaries; separate PDF median 7.5 pp. (22 papers) |
| Theory & formal analysis | 7 | 5 / 2 | 8.11 | 2.621/7 attached boundaries; separate PDF median 13 pp. (7 papers) |
| Empirical analysis, diagnosis & auditing | 10 | 4.5 / 3 | 8.09 | Not resolved0/10 attached boundaries; separate PDF median 5 pp. (9 papers) |
| Datasets, benchmarks & measurement | 20 | 6 / 4.5 | 8.11 | 5.962/20 attached boundaries; separate PDF median 11 pp. (20 papers) |
| Efficiency, inference & systems | 19 | 7 / 4 | 8.09 | Not resolved0/19 attached boundaries; separate PDF median 6 pp. (15 papers) |
| Domain-grounded methods & physical systems | 12 | 5.5 / 2.5 | 8.05 | Not resolved0/12 attached boundaries; separate PDF median 6 pp. (11 papers) |
Search all 141 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 141 of 141 records.
| Paper / category | PDF pages | Main F / T | Appendix F / T | Body / appendix span ≈ pp. | Notes |
|---|---|---|---|---|---|
| Black-box Membership Inference Attacks on the Pre-training Data of Image-generation ModelsEmpirical analysis, diagnosis & auditing | 10 | 7 / 1 | 0 / 0 | 8.14 / Unresolved | No attached appendix boundary detectedSeparate files: 3 pp.; 2 figures / 2 tables; excluded from main totals |
| A Style is Worth One Code: Unlocking Code-to-Style Image Generation with Discrete Style SpaceArchitecture & representation | 10 | 8 / 3 | 0 / 0 | 8.08 / Unresolved | No attached appendix boundary detectedSeparate files: 8 pp.; 6 figures / 2 tables; excluded from main totals |
| Advancing Image Classification with Discrete Diffusion Classification ModelingTraining, objectives & optimization | 11 | 4 / 1 | 0 / 0 | 8.08 / Unresolved | No attached appendix boundary detectedSeparate files: 7 pp.; 2 figures / 5 tables; excluded from main totals |
| Customized Fusion: A Closed-Loop Dynamic Network for Adaptive Multi-Task-Aware Infrared-Visible Image FusionArchitecture & representation | 11 | 8 / 5 | 0 / 0 | 8.13 / Unresolved | No attached appendix boundary detectedSeparate files: 5 pp.; 4 figures / 8 tables; excluded from main totals |
| Dual Band Thermal Videography: Separating Time-Varying Reflection and Emission Near Ambient ConditionsDomain-grounded methods & physical systems | 10 | 6 / 2 | 0 / 0 | 8.08 / Unresolved | No attached appendix boundary detectedSeparate files: 2 pp.; 5 figures / 0 tables; excluded from main totals |
| Data Leakage Detection and De-duplication in Large Scale Geospatial Image DatasetsEmpirical analysis, diagnosis & auditing | 10 | 4 / 4 | 0 / 0 | 8.13 / Unresolved | No attached appendix boundary detectedSeparate files: 5 pp.; 5 figures / 2 tables; excluded from main totals |
| Does YOLO Really Need to See Every Training Image in Every Epoch?Efficiency, inference & systems | 10 | 4 / 5 | 0 / 0 | 8.09 / Unresolved | No attached appendix boundary detected |
| ANTS: Adaptive Negative Textual Space Shaping for OOD Detection via Test-Time MLLM Understanding and ReasoningArchitecture & representation | 11 | 7 / 6 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 8 pp.; 4 figures / 7 tables; excluded from main totals |
| Fine-grained Image Aesthetic Assessment: Learning Discriminative Scores from Relative RanksDatasets, benchmarks & measurement | 11 | 5 / 5 | 0 / 0 | 8.13 / Unresolved | No attached appendix boundary detectedSeparate files: 5 pp.; 6 figures / 3 tables; excluded from main totals |
| RAVEN: Erasing Invisible Watermarks via Novel View SynthesisEmpirical analysis, diagnosis & auditing | 10 | 5 / 6 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSource correction retainedSeparate files: 5 pp.; 4 figures / 1 tables; excluded from main totals |
| MetaSpectra+: A Compact Broadband Metasurface Camera for Snapshot Hyperspectral+ ImagingDomain-grounded methods & physical systems | 11 | 5 / 1 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 5 pp.; 6 figures / 2 tables; excluded from main totals |
| ARGUS: Defending Against Multimodal Indirect Prompt Injection via Steering Instruction-Following BehaviorEmpirical analysis, diagnosis & auditing | 10 | 3 / 1 | 0 / 0 | 8.14 / Unresolved | No attached appendix boundary detectedSeparate files: 5 pp.; 4 figures / 2 tables; excluded from main totals |
| NuWa: Deriving Lightweight Class-Specific Vision Transformers for Edge DevicesEfficiency, inference & systems | 11 | 12 / 4 | 0 / 0 | 8.09 / Unresolved | No attached appendix boundary detectedSeparate files: 6 pp.; 3 figures / 5 tables; excluded from main totals |
| LDP-Slicing: Local Differential Privacy for Images via Randomized Bit-Plane SlicingTheory & formal analysis | 10 | 8 / 5 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 4 pp.; 2 figures / 2 tables; excluded from main totals |
| TEAR: Temporal-aware Automated Red-teaming for Text-to-Video ModelsEmpirical analysis, diagnosis & auditing | 10 | 7 / 3 | 0 / 0 | 8.13 / Unresolved | No attached appendix boundary detected |
| Towards Photorealistic and Efficient Bokeh Rendering via Diffusion FrameworkArchitecture & representation | 10 | 7 / 2 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 5 pp.; 6 figures / 2 tables; excluded from main totals |
| Plant Taxonomy Meets Plant Counting: A Fine-Grained, Taxonomic Dataset for Counting Hundreds of Plant SpeciesDatasets, benchmarks & measurement | 11 | 6 / 4 | 0 / 0 | 8.08 / Unresolved | No attached appendix boundary detectedSeparate files: 4 pp.; 6 figures / 0 tables; excluded from main totals |
| NOWA: Null-space Optical Watermark for Invisible Capture Fingerprinting and Tamper LocalizationDomain-grounded methods & physical systems | 11 | 5 / 2 | 0 / 0 | 8.1 / Unresolved | No attached appendix boundary detectedSeparate files: 4 pp.; 3 figures / 0 tables; excluded from main totals |
| ViT^3: Unlocking Test-Time Training in VisionEmpirical analysis, diagnosis & auditing | 11 | 4 / 10 | 0 / 0 | 8.11 / Unresolved | No attached appendix boundary detectedSeparate files: 3 pp.; 0 figures / 3 tables; excluded from main totals |
| UnReflectAnything: RGB-Only Highlight Removal by Rendering Synthetic Specular SupervisionArchitecture & representation | 10 | 6 / 3 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 4 pp.; 3 figures / 2 tables; excluded from main totals |
| Revisiting Geometric Obfuscation with Dual Convergent Lines for Privacy-Preserving Image Queries in Visual LocalizationArchitecture & representation | 11 | 7 / 6 | 0 / 0 | 8.11 / Unresolved | No attached appendix boundary detectedSeparate files: 9 pp.; 4 figures / 2 tables; excluded from main totals |
| Rethinking Dataset Distillation: Hard Truths about Soft LabelsEmpirical analysis, diagnosis & auditing | 10 | 3 / 3 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 17 pp.; 10 figures / 11 tables; excluded from main totals |
| Hearing the Room Through the Shape of the Drum: Modal-Guided Sound Recovery from Multi-Point Surface VibrationsDomain-grounded methods & physical systems | 10 | 7 / 0 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 4 pp.; 1 figures / 4 tables; excluded from main totals |
| 4D Primitive-Mache: Glueing Primitives for Persistent 4D Scene ReconstructionArchitecture & representation | 10 | 5 / 2 | 0 / 0 | 8.05 / 1.91 | No automatic flagsSeparate files: 3 pp.; 2 figures / 3 tables; excluded from main totals |
| 3DReflecNet: A Large-Scale Dataset for 3D Reconstruction of Reflective, Transparent, and Low-Texture ObjectsDatasets, benchmarks & measurement | 12 | 12 / 5 | 0 / 0 | 8.17 / Unresolved | No attached appendix boundary detectedSeparate files: 18 pp.; 16 figures / 9 tables; excluded from main totals |
| MAMMA: Markerless Accurate Multi-person Motion AcquisitionArchitecture & representation | 12 | 7 / 5 | 0 / 0 | 8.15 / Unresolved | No attached appendix boundary detectedSeparate files: 10 pp.; 15 figures / 5 tables; excluded from main totals |
| Energy-GS: Image Energy-guided Pose Alignment Gaussian Splatting with redesigned pose gradient flowTraining, objectives & optimization | 10 | 7 / 4 | 0 / 0 | 8.1 / Unresolved | No attached appendix boundary detectedSeparate files: 6 pp.; 2 figures / 2 tables; excluded from main totals |
| Natural Human Motion Recovery by Aligning High-Order Temporal Dynamics from Monocular VideosArchitecture & representation | 10 | 4 / 5 | 0 / 0 | 8.12 / Unresolved | No attached appendix boundary detectedSeparate files: 3 pp.; 2 figures / 3 tables; excluded from main totals |
| GLINT: Modeling Scene-Scale Transparency via Gaussian Radiance TransportArchitecture & representation | 10 | 8 / 3 | 0 / 0 | 8.18 / Unresolved | No attached appendix boundary detectedSeparate files: 9 pp.; 9 figures / 1 tables; excluded from main totals |
| Efficiently Reconstructing Dynamic Scenes One D4RT at a TimeEfficiency, inference & systems | 11 | 6 / 9 | 0 / 0 | 8.26 / Unresolved | No attached appendix boundary detectedSeparate files: 6 pp.; 7 figures / 3 tables; excluded from main totals |
| MeshSplatting: Differentiable Rendering with Opaque MeshesArchitecture & representation | 10 | 7 / 4 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 6 pp.; 7 figures / 7 tables; excluded from main totals |
| Neural Field-Based 3D Surface Reconstruction of Microstructures from Multi-Detector Signals in Scanning Electron MicroscopyDomain-grounded methods & physical systems | 12 | 6 / 1 | 0 / 0 | 8.15 / Unresolved | No attached appendix boundary detectedSeparate files: 9 pp.; 8 figures / 0 tables; excluded from main totals |
| PoseGAM: Robust Unseen Object Pose Estimation via Geometry-Aware Multi-View ReasoningArchitecture & representation | 12 | 4 / 5 | 0 / 0 | 8.1 / Unresolved | No attached appendix boundary detectedSeparate files: 7 pp.; 3 figures / 5 tables; excluded from main totals |
| FUSER: Feed-Forward Multiview 3D Registration Transformer and SE(3)^N Diffusion RefinementArchitecture & representation | 11 | 5 / 5 | 0 / 0 | 8.11 / Unresolved | No attached appendix boundary detectedSeparate files: 3 pp.; 1 figures / 0 tables; excluded from main totals |
| Proxy-GS: Unified Occlusion Priors for Training and Inference in Structured 3D Gaussian SplattingEfficiency, inference & systems | 10 | 6 / 4 | 0 / 0 | 8.06 / Unresolved | No attached appendix boundary detectedSeparate files: 4 pp.; 3 figures / 6 tables; excluded from main totals |
| PhyGaP: Physically-Grounded Gaussians with Polarization CuesDomain-grounded methods & physical systems | 11 | 9 / 2 | 0 / 0 | 8.11 / Unresolved | No attached appendix boundary detectedSeparate files: 7 pp.; 7 figures / 3 tables; excluded from main totals |
| SAM 3D Body: Robust Full-Body Human Mesh RecoveryArchitecture & representation | 11 | 5 / 5 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 5 pp.; 5 figures / 3 tables; excluded from main totals |
| Residual Primitive Fitting of 3D Shapes with SuperFrustaArchitecture & representation | 10 | 8 / 4 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 22 pp.; 13 figures / 9 tables; excluded from main totals |
| RetimeGS: Continuous-Time Reconstruction of 4D Gaussian SplattingArchitecture & representation | 11 | 7 / 2 | 0 / 0 | 8.16 / Unresolved | No attached appendix boundary detectedSeparate files: 4 pp.; 3 figures / 5 tables; excluded from main totals |
| PPISP: Physically-Plausible Compensation and Control of Photometric Variations in Radiance Field ReconstructionDomain-grounded methods & physical systems | 10 | 4 / 5 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 7 pp.; 6 figures / 4 tables; excluded from main totals |
| Selfi: Self-improving Reconstruction Engine via 3D Geometric Feature AlignmentTraining, objectives & optimization | 11 | 6 / 6 | 0 / 0 | 8.11 / Unresolved | No attached appendix boundary detectedSeparate files: 8 pp.; 7 figures / 3 tables; excluded from main totals |
| SAM 3D: 3Dfy Anything in ImagesArchitecture & representation | 13 | 9 / 4 | 0 / 0 | 8.04 / Unresolved | No attached appendix boundary detectedSeparate files: 19 pp.; 12 figures / 8 tables; excluded from main totals |
| SmokeSVD: Smoke Reconstruction from A Single View via Progressive Novel View Synthesis and Refinement with Diffusion ModelsArchitecture & representation | 11 | 16 / 6 | 0 / 0 | 8.15 / Unresolved | No attached appendix boundary detectedSeparate files: 10 pp.; 17 figures / 4 tables; excluded from main totals |
| SparseWorld-TC: Trajectory-Conditioned Sparse Occupancy World ModelArchitecture & representation | 10 | 6 / 4 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 5 pp.; 7 figures / 7 tables; excluded from main totals |
| Z-Order Transformer for Feed-Forward Gaussian SplattingEfficiency, inference & systems | 10 | 7 / 6 | 0 / 0 | 8.13 / Unresolved | No attached appendix boundary detectedSeparate files: 4 pp.; 0 figures / 4 tables; excluded from main totals |
| SeeGroup: Multi-Layer Depth Estimation of Transparent Surfaces via Self-Determined GroupingTraining, objectives & optimization | 11 | 5 / 4 | 0 / 0 | 8.08 / Unresolved | No attached appendix boundary detectedSeparate files: 4 pp.; 1 figures / 1 tables; excluded from main totals |
| SPARK: Sim-ready Part-level Articulated Reconstruction with VLM KnowledgeArchitecture & representation | 11 | 7 / 4 | 0 / 0 | 8.13 / Unresolved | No attached appendix boundary detectedSeparate files: 6 pp.; 2 figures / 1 tables; excluded from main totals |
| 3D-LATTE: Latent Space 3D Editing from Textual InstructionsArchitecture & representation | 10 | 9 / 2 | 0 / 0 | 8.11 / Unresolved | No attached appendix boundary detectedSeparate files: 9 pp.; 6 figures / 4 tables; excluded from main totals |
| Breaking Semantic Boundaries: Distribution-Guided Semantic Exploration for Creative GenerationTraining, objectives & optimization | 10 | 8 / 4 | 0 / 0 | 8.13 / Unresolved | No attached appendix boundary detectedSeparate files: 7 pp.; 6 figures / 0 tables; excluded from main totals |
| Differentiable Vector Quantization for Rate-Distortion Optimization of Generative Image CompressionTraining, objectives & optimization | 11 | 10 / 1 | 0 / 0 | 8.1 / Unresolved | No attached appendix boundary detectedSeparate files: 12 pp.; 14 figures / 1 tables; excluded from main totals |
| ComPose: A Unified Completion-Pose Framework for Robust Category-Level Object Pose EstimationArchitecture & representation | 10 | 5 / 7 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detected |
| AnchorFlow: Training-Free 3D Editing via Latent Anchor-Aligned FlowsArchitecture & representation | 11 | 8 / 2 | 0 / 0 | 8.09 / Unresolved | No attached appendix boundary detectedSeparate files: 10 pp.; 4 figures / 1 tables; excluded from main totals |
| Guiding a Diffusion Model by Swapping Its TokensArchitecture & representation | 10 | 8 / 7 | 0 / 0 | 8.1 / Unresolved | No attached appendix boundary detectedSeparate files: 7 pp.; 9 figures / 6 tables; excluded from main totals |
| FINER: MLLMs Hallucinate under Fine-grained Negative QueriesDatasets, benchmarks & measurement | 10 | 5 / 5 | 0 / 0 | 8.22 / Unresolved | No attached appendix boundary detectedSeparate files: 27 pp.; 20 figures / 12 tables; excluded from main totals |
| CoSMo3D: Open-World Promptable 3D Semantic Segmentation through LLM-Guided Canonical Spatial ModelingArchitecture & representation | 10 | 6 / 2 | 0 / 0 | 8.09 / Unresolved | No attached appendix boundary detectedSeparate files: 4 pp.; 5 figures / 3 tables; excluded from main totals |
| PixelDiT: Pixel Diffusion Transformers for Image GenerationArchitecture & representation | 10 | 5 / 5 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 17 pp.; 22 figures / 12 tables; excluded from main totals |
| ChordEdit: One-Step Low-Energy Transport for Image EditingEfficiency, inference & systems | 10 | 11 / 3 | 0 / 0 | 8.04 / Unresolved | No attached appendix boundary detectedSeparate files: 24 pp.; 13 figures / 2 tables; excluded from main totals |
| MDCS-MoAME: Multi-directional Composite Scanning with Mixture of Attention and Mamba Experts for Cancer Survival PredictionDomain-grounded methods & physical systems | 10 | 6 / 5 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detected |
| GeoViS: Geospatially Rewarded Visual Search for Remote Sensing Visual GroundingArchitecture & representation | 11 | 4 / 5 | 0 / 0 | 8.11 / Unresolved | No attached appendix boundary detectedSeparate files: 8 pp.; 0 figures / 4 tables; excluded from main totals |
| RobotSeg: A Model and Dataset for Segmenting Robots in Image and VideoDatasets, benchmarks & measurement | 11 | 9 / 4 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 14 pp.; 11 figures / 4 tables; excluded from main totals |
| Faithful Contouring: Near-Lossless 3D Voxel Representation Free from Iso-surfaceArchitecture & representation | 11 | 7 / 2 | 0 / 0 | 8.08 / Unresolved | No attached appendix boundary detectedSeparate files: 2 pp.; 4 figures / 0 tables; excluded from main totals |
| SeaCache: Spectral-Evolution-Aware Cache for Accelerating Diffusion ModelsEfficiency, inference & systems | 12 | 10 / 4 | 0 / 0 | 8.12 / Unresolved | No attached appendix boundary detectedSeparate files: 7 pp.; 2 figures / 6 tables; excluded from main totals |
| PAS: A Training-Free Stabilizer for Temporal Encoding in Video LLMsTheory & formal analysis | 10 | 5 / 2 | 0 / 0 | 8.15 / Unresolved | No attached appendix boundary detectedSeparate files: 3 pp.; 0 figures / 0 tables; excluded from main totals |
| PAVAS: Physics-Aware Video-to-Audio SynthesisDomain-grounded methods & physical systems | 11 | 3 / 3 | 0 / 0 | 8.19 / Unresolved | No attached appendix boundary detectedSeparate files: 14 pp.; 6 figures / 8 tables; excluded from main totals |
| SenCache: Accelerating Diffusion Model Inference via Sensitivity-Aware CachingEfficiency, inference & systems | 10 | 4 / 3 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 1 pp.; 1 figures / 0 tables; excluded from main totals |
| S^2AM3D: Scale-controllable Part Segmentation of 3D Point CloudsArchitecture & representation | 10 | 7 / 3 | 0 / 0 | 8.1 / Unresolved | No attached appendix boundary detectedSeparate files: 3 pp.; 2 figures / 3 tables; excluded from main totals |
| Native and Compact Structured Latents for 3D GenerationArchitecture & representation | 11 | 8 / 3 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 15 pp.; 6 figures / 4 tables; excluded from main totals |
| Streaming Diffusion Model for Fast Infrared and Visible Video FusionEfficiency, inference & systems | 10 | 9 / 3 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detected |
| Scalable Multi-View Subspace Clustering with Tensorized Anchor GuidanceEfficiency, inference & systems | 10 | 6 / 4 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 6 pp.; 4 figures / 4 tables; excluded from main totals |
| SliderEdit: Continuous Image Editing with Fine-Grained Instruction ControlArchitecture & representation | 10 | 7 / 2 | 0 / 0 | 8.13 / Unresolved | No attached appendix boundary detectedSeparate files: 9 pp.; 10 figures / 0 tables; excluded from main totals |
| ProPhy: Progressive Physical Alignment for Dynamic World SimulationArchitecture & representation | 10 | 8 / 4 | 0 / 0 | 8.12 / Unresolved | No attached appendix boundary detectedSeparate files: 6 pp.; 6 figures / 0 tables; excluded from main totals |
| Breaking the Scalability Limit of Multi-Projector Calibration with Embedded CamerasEfficiency, inference & systems | 10 | 12 / 1 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 6 pp.; 9 figures / 0 tables; excluded from main totals |
| INSID3: Training-Free In-Context Segmentation with DINOv3Architecture & representation | 11 | 7 / 3 | 0 / 0 | 8.25 / Unresolved | No attached appendix boundary detectedSeparate files: 8 pp.; 3 figures / 10 tables; excluded from main totals |
| CodeV: Code with Images for Faithful Visual Reasoning via Tool-Aware Policy OptimizationTraining, objectives & optimization | 11 | 6 / 2 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 13 pp.; 4 figures / 6 tables; excluded from main totals |
| Chorus: Multi-Teacher Pretraining for Holistic 3D Gaussian Scene EncodingTraining, objectives & optimization | 12 | 8 / 10 | 0 / 0 | 8.16 / Unresolved | No attached appendix boundary detectedSeparate files: 10 pp.; 9 figures / 3 tables; excluded from main totals |
| GaussianFluent: Gaussian Simulation for Dynamic Scenes with Mixed MaterialsDomain-grounded methods & physical systems | 11 | 7 / 3 | 0 / 0 | 8.04 / Unresolved | No attached appendix boundary detectedSeparate files: 6 pp.; 7 figures / 1 tables; excluded from main totals |
| MARCO: Navigating the Unseen Space of Semantic CorrespondenceTraining, objectives & optimization | 10 | 5 / 4 | 0 / 0 | 8.17 / Unresolved | No attached appendix boundary detectedSeparate files: 8 pp.; 2 figures / 12 tables; excluded from main totals |
| Featurising Pixels from Dynamic 3D Scenes with Linear In-Context LearnersTraining, objectives & optimization | 10 | 8 / 5 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 2 pp.; 0 figures / 1 tables; excluded from main totals |
| NitroGen: An Open Foundation Model for Generalist Gaming AgentsDatasets, benchmarks & measurement | 11 | 7 / 0 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 3 pp.; 2 figures / 0 tables; excluded from main totals |
| PAI-Bench: A Comprehensive Benchmark For Physical AIDatasets, benchmarks & measurement | 15 | 7 / 6 | 0 / 0 | 8.17 / Unresolved | No attached appendix boundary detectedSource correction retainedSeparate files: 17 pp.; 15 figures / 3 tables; excluded from main totals |
| From Pairs to Sequences: Track-Aware Policy Gradients for Keypoint DetectionTraining, objectives & optimization | 11 | 5 / 5 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 6 pp.; 5 figures / 4 tables; excluded from main totals |
| PR-MaGIC: Prompt Refinement Via Mask Decoder Gradient Flow For In-Context SegmentationTraining, objectives & optimization | 10 | 7 / 2 | 0 / 0 | 8.08 / Unresolved | No attached appendix boundary detectedSeparate files: 11 pp.; 9 figures / 4 tables; excluded from main totals |
| InfiniBench: Infinite Benchmarking for Visual Spatial Reasoning with Customizable Scene ComplexityDatasets, benchmarks & measurement | 12 | 13 / 6 | 0 / 0 | 8.11 / 3.11 | No automatic flagsSeparate files: 9 pp.; 12 figures / 2 tables; excluded from main totals |
| Linear Fundamental Matrix Estimation from 7 or 5 PointsTheory & formal analysis | 10 | 5 / 1 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 4 pp.; 2 figures / 1 tables; excluded from main totals |
| MAGICIAN: Efficient Long-Term Planning with Imagined Gaussians for Active MappingEfficiency, inference & systems | 10 | 8 / 5 | 0 / 0 | 8.12 / Unresolved | No attached appendix boundary detectedSeparate files: 8 pp.; 3 figures / 4 tables; excluded from main totals |
| RefAV: Towards Planning-Centric Scenario MiningDatasets, benchmarks & measurement | 12 | 4 / 4 | 0 / 0 | 8.04 / Unresolved | No attached appendix boundary detectedSeparate files: 16 pp.; 4 figures / 3 tables; excluded from main totals |
| R2-Seg: Training-Free OOD Medical Tumor Segmentation via Anatomical Reasoning and Statistical RejectionDomain-grounded methods & physical systems | 10 | 5 / 3 | 0 / 0 | 8.04 / Unresolved | No attached appendix boundary detectedSeparate files: 9 pp.; 3 figures / 4 tables; excluded from main totals |
| Memory-Augmented Scene Understanding and Exploration for Open-World Aerial Object-Goal NavigationArchitecture & representation | 11 | 3 / 5 | 0 / 0 | 8.09 / Unresolved | No attached appendix boundary detectedSeparate files: 5 pp.; 4 figures / 0 tables; excluded from main totals |
| OccuFly: A 3D Vision Benchmark for Semantic Scene Completion from the Aerial PerspectiveDatasets, benchmarks & measurement | 12 | 5 / 5 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 7 pp.; 6 figures / 7 tables; excluded from main totals |
| The SA-FARI Dataset: Segment Anything in Footage of Animals for Recognition and IdentificationDatasets, benchmarks & measurement | 11 | 5 / 5 | 0 / 0 | 8.34 / Unresolved | No attached appendix boundary detectedSeparate files: 3 pp.; 0 figures / 0 tables; excluded from main totals |
| SoccerMaster: A Vision Foundation Model for Soccer UnderstandingArchitecture & representation | 12 | 3 / 7 | 0 / 0 | 8.09 / Unresolved | No attached appendix boundary detectedSeparate files: 11 pp.; 3 figures / 5 tables; excluded from main totals |
| VGGT-Segmentor: Geometry-Enhanced Cross-View SegmentationArchitecture & representation | 11 | 4 / 8 | 0 / 0 | 8.16 / Unresolved | No attached appendix boundary detectedSeparate files: 4 pp.; 6 figures / 1 tables; excluded from main totals |
| VS-Bench: Evaluating VLMs for Strategic Abilities in Multi-Agent EnvironmentsDatasets, benchmarks & measurement | 12 | 6 / 4 | 0 / 0 | 8.1 / Unresolved | No attached appendix boundary detectedSeparate files: 24 pp.; 23 figures / 7 tables; excluded from main totals |
| VGGT-ΩArchitecture & representation | 14 | 4 / 3 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 3 pp.; 0 figures / 0 tables; excluded from main totals |
| Monocular Open Vocabulary Occupancy Prediction for Indoor ScenesTraining, objectives & optimization | 11 | 5 / 4 | 0 / 0 | 8.09 / Unresolved | No attached appendix boundary detectedSeparate files: 2 pp.; 4 figures / 2 tables; excluded from main totals |
| Evidential Neural Radiance FieldsTraining, objectives & optimization | 10 | 8 / 2 | 0 / 0 | 8.08 / Unresolved | No attached appendix boundary detectedSource correction retainedSeparate files: 12 pp.; 6 figures / 5 tables; excluded from main totals |
| BoostSLT: Boosting Sign Language Translation via a Plug-and-Play Diffusion-Based Semantic EnhancerArchitecture & representation | 10 | 4 / 3 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 3 pp.; 4 figures / 3 tables; excluded from main totals |
| AT-VLA: Adaptive Tactile Injection for Enhanced Feedback Reaction in Vision-Language-Action ModelsArchitecture & representation | 11 | 4 / 3 | 0 / 0 | 8.09 / Unresolved | No attached appendix boundary detected |
| AToken: A Unified Tokenizer for VisionArchitecture & representation | 11 | 7 / 5 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 18 pp.; 8 figures / 10 tables; excluded from main totals |
| ImmerIris: A Large-Scale Dataset and Benchmark for Off-Axis and Unconstrained Iris Recognition in Immersive ApplicationsDatasets, benchmarks & measurement | 10 | 6 / 7 | 0 / 0 | 8.14 / Unresolved | No attached appendix boundary detectedSeparate files: 6 pp.; 5 figures / 5 tables; excluded from main totals |
| Global-Aware Edge Prioritization for Pose Graph InitializationEfficiency, inference & systems | 10 | 4 / 3 | 0 / 0 | 8.11 / Unresolved | No attached appendix boundary detectedSeparate files: 2 pp.; 0 figures / 0 tables; excluded from main totals |
| Confusion-Aware Spectral Regularizer for Long-Tailed RecognitionTraining, objectives & optimization | 11 | 3 / 7 | 0 / 0 | 8.16 / Unresolved | No attached appendix boundary detected |
| Learning Diffeomorphism for Medical Image Registration with Time-Embedded Architectures Using Semigroup RegularizationTraining, objectives & optimization | 11 | 8 / 7 | 0 / 0 | 8.1 / Unresolved | No attached appendix boundary detectedSource correction retainedSeparate files: 10 pp.; 8 figures / 7 tables; excluded from main totals |
| QuadSync: Quadrifocal Tensor Synchronization via Tucker DecompositionTheory & formal analysis | 10 | 3 / 0 | 0 / 0 | 8.16 / Unresolved | No attached appendix boundary detectedSeparate files: 13 pp.; 3 figures / 7 tables; excluded from main totals |
| OLATverse: A Large-scale Real-world Object Dataset with Precise Lighting ControlDatasets, benchmarks & measurement | 12 | 7 / 3 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 13 pp.; 13 figures / 1 tables; excluded from main totals |
| Learning Latent Concepts for Detecting Out-of-Distribution ObjectsArchitecture & representation | 11 | 5 / 6 | 0 / 0 | 8.13 / Unresolved | No attached appendix boundary detectedSeparate files: 3 pp.; 4 figures / 3 tables; excluded from main totals |
| Molmo2: Open Weights and Data for Vision-Language Models with Video Understanding and GroundingDatasets, benchmarks & measurement | 17 | 3 / 7 | 1 / 0 | 8.18 / 8.8 | No automatic flagsSeparate files: 31 pp.; 27 figures / 14 tables; excluded from main totals |
| OpenDance: Multimodal Controllable 3D Dance Generation with Large-scale Internet DataDatasets, benchmarks & measurement | 11 | 5 / 9 | 0 / 0 | 8.08 / Unresolved | No attached appendix boundary detectedSeparate files: 8 pp.; 7 figures / 6 tables; excluded from main totals |
| Optical Flow Matching: Reframing Optical Flow as Continuous Transport DynamicsTraining, objectives & optimization | 10 | 5 / 4 | 0 / 0 | 8.05 / 1.07 | No automatic flags |
| Learning Like Humans: Analogical Concept Learning for Generalized Category DiscoveryArchitecture & representation | 10 | 4 / 4 | 0 / 0 | 8.13 / Unresolved | No attached appendix boundary detectedSeparate files: 5 pp.; 4 figures / 4 tables; excluded from main totals |
| SocialNav: Training Human-Inspired Foundation Model for Socially-Aware Embodied NavigationArchitecture & representation | 11 | 4 / 4 | 0 / 0 | 8.11 / Unresolved | No attached appendix boundary detectedSeparate files: 8 pp.; 5 figures / 5 tables; excluded from main totals |
| POLAR: A Portrait OLAT Dataset and Generative Framework for Illumination-Aware Face ModelingDatasets, benchmarks & measurement | 11 | 9 / 2 | 0 / 0 | 8.1 / Unresolved | No attached appendix boundary detectedSeparate files: 8 pp.; 13 figures / 1 tables; excluded from main totals |
| Structural Action Transformer for 3D Dexterous ManipulationArchitecture & representation | 12 | 8 / 6 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 6 pp.; 2 figures / 3 tables; excluded from main totals |
| SEATrack: Simple, Efficient, and Adaptive Multimodal TrackerEfficiency, inference & systems | 11 | 7 / 6 | 0 / 0 | 8.11 / Unresolved | No attached appendix boundary detected |
| Understanding and Enforcing Weight Disentanglement in Task ArithmeticTheory & formal analysis | 10 | 6 / 2 | 0 / 0 | 8.11 / Unresolved | No attached appendix boundary detectedSeparate files: 20 pp.; 5 figures / 4 tables; excluded from main totals |
| Understanding Task Transfer in Vision-Language ModelsEmpirical analysis, diagnosis & auditing | 10 | 10 / 2 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 17 pp.; 25 figures / 2 tables; excluded from main totals |
| U^2Flow: Uncertainty-Aware Unsupervised Optical Flow EstimationTraining, objectives & optimization | 11 | 5 / 6 | 0 / 0 | 8.13 / Unresolved | No attached appendix boundary detectedSeparate files: 7 pp.; 8 figures / 4 tables; excluded from main totals |
| Relightable Holoported Characters: Capturing and Relighting Dynamic Human Performance from Sparse ViewsArchitecture & representation | 11 | 7 / 2 | 0 / 0 | 8.08 / Unresolved | No attached appendix boundary detectedSource correction retainedSeparate files: 9 pp.; 10 figures / 5 tables; excluded from main totals |
| TESO: Online Tracking of Essential Matrix by Stochastic OptimizationDomain-grounded methods & physical systems | 9 | 3 / 4 | 0 / 0 | 7.76 / Unresolved | No attached appendix boundary detectedSeparate files: 5 pp.; 3 figures / 1 tables; excluded from main totals |
| Differentiable Laplacian Matrix Guided Superpixel SegmentationTraining, objectives & optimization | 11 | 6 / 3 | 0 / 0 | 8.1 / Unresolved | No attached appendix boundary detectedSeparate files: 6 pp.; 8 figures / 2 tables; excluded from main totals |
| CURE: Curriculum-guided Multi-task Training for Reliable Anatomy Grounded Report GenerationTraining, objectives & optimization | 11 | 3 / 7 | 0 / 0 | 8.14 / Unresolved | No attached appendix boundary detectedSeparate files: 20 pp.; 4 figures / 13 tables; excluded from main totals |
| Efficient Unrolled Networks for Large-Scale 3D Inverse ProblemsEfficiency, inference & systems | 11 | 4 / 4 | 0 / 0 | 8.13 / Unresolved | No attached appendix boundary detectedSeparate files: 4 pp.; 6 figures / 0 tables; excluded from main totals |
| CineBrain: A Large-Scale Multi-Modal Audiovisual Brain Dataset for Brain-Conditioned Video GenerationDatasets, benchmarks & measurement | 11 | 6 / 3 | 0 / 0 | 8.12 / Unresolved | No attached appendix boundary detectedSeparate files: 6 pp.; 7 figures / 2 tables; excluded from main totals |
| DK-DDIL: Adaptive Knowledge Retention for Dynamic Domain-Incremental Learning in Medical ImagingTraining, objectives & optimization | 10 | 3 / 1 | 0 / 0 | 8.1 / Unresolved | No attached appendix boundary detectedSeparate files: 5 pp.; 2 figures / 2 tables; excluded from main totals |
| Spectrum from Defocus: Fast Spectral Imaging with Chromatic Focal StackEfficiency, inference & systems | 11 | 5 / 1 | 0 / 0 | 8.1 / Unresolved | No attached appendix boundary detectedSeparate files: 16 pp.; 20 figures / 4 tables; excluded from main totals |
| FedAdamom: Adaptive Momentum for Improved Generalization in Federated OptimizationTheory & formal analysis | 11 | 3 / 3 | 0 / 0 | 8.13 / Unresolved | No attached appendix boundary detectedSeparate files: 15 pp.; 1 figures / 6 tables; excluded from main totals |
| FILTR: Extracting Topological Features from Pretrained 3D ModelsEmpirical analysis, diagnosis & auditing | 11 | 8 / 3 | 0 / 0 | 8.08 / Unresolved | No attached appendix boundary detectedSeparate files: 7 pp.; 10 figures / 9 tables; excluded from main totals |
| Dual-level Adapter Boosting Prompt-free Curvilinear Structure SegmentationArchitecture & representation | 11 | 10 / 5 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detected |
| SDTrack: A Baseline for Event-based Tracking via Spiking Neural NetworksEfficiency, inference & systems | 10 | 8 / 4 | 0 / 0 | 8.11 / Unresolved | No attached appendix boundary detected |
| Learning Convex Decomposition via Feature FieldsTraining, objectives & optimization | 11 | 11 / 2 | 0 / 0 | 8.11 / Unresolved | No attached appendix boundary detectedSeparate files: 3 pp.; 0 figures / 2 tables; excluded from main totals |
| SimScale: Learning to Drive via Real-World Simulation at ScaleDatasets, benchmarks & measurement | 10 | 5 / 4 | 0 / 0 | 8.16 / Unresolved | No attached appendix boundary detectedSeparate files: 13 pp.; 9 figures / 6 tables; excluded from main totals |
| LATA: Laplacian-Assisted Transductive Adaptation for Conformal Uncertainty in Medical VLMsTraining, objectives & optimization | 10 | 4 / 3 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 9 pp.; 5 figures / 9 tables; excluded from main totals |
| Thinking with Drafts: Speculative Temporal Reasoning for Efficient Long Video UnderstandingEfficiency, inference & systems | 12 | 5 / 5 | 0 / 0 | 8.08 / Unresolved | No attached appendix boundary detectedSeparate files: 7 pp.; 8 figures / 3 tables; excluded from main totals |
| Texvent: Asynchronous Event Data Simulation via Text PromptEfficiency, inference & systems | 10 | 7 / 3 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 7 pp.; 7 figures / 7 tables; excluded from main totals |
| Learning Eigenstructures of Unstructured Data ManifoldsTheory & formal analysis | 14 | 9 / 1 | 0 / 0 | 8.05 / 2.62 | No automatic flagsSeparate files: 27 pp.; 20 figures / 12 tables; excluded from main totals |
| Medic-AD: Towards Medical Vision-Language Model's Clinical IntelligenceArchitecture & representation | 11 | 5 / 5 | 0 / 0 | 8.18 / Unresolved | No attached appendix boundary detectedSeparate files: 2 pp.; 1 figures / 5 tables; excluded from main totals |
| Wan-Weaver: Interleaved Multi-modal Generation via Decoupled TrainingTraining, objectives & optimization | 12 | 8 / 3 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 18 pp.; 7 figures / 8 tables; excluded from main totals |
| WorldLens: Full-Spectrum Evaluations of Driving World Models in Real WorldDatasets, benchmarks & measurement | 15 | 8 / 3 | 0 / 0 | 8.17 / Unresolved | No attached appendix boundary detectedSeparate files: 62 pp.; 29 figures / 27 tables; excluded from main totals |
| Mapping NetworksArchitecture & representation | 9 | 5 / 8 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 6 pp.; 1 figures / 12 tables; excluded from main totals |
| SegMoTE: Token-Level Mixture of Experts for Medical Image SegmentationArchitecture & representation | 11 | 7 / 4 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 6 pp.; 5 figures / 4 tables; excluded from main totals |
| Adversarial Style Optimization: Enhancing VLM Jailbreaks by GRPO-based Stylistic Triggers OptimizationEmpirical analysis, diagnosis & auditing | 9 | 2 / 4 | 0 / 0 | 8.05 / Unresolved | No attached appendix boundary detectedSeparate files: 5 pp.; 4 figures / 2 tables; excluded from main totals |
No matching papers. Reset or use a shorter query.
Measurement method
Section titled “Measurement method”The offline extractor uses Poppler page coordinates, caption labels and top-level headings. CVPR’s title-case numbered headings and two-column layout are treated separately from ICLR’s single-column conventions. For two-column pages, the left and right text columns each occupy half a page-equivalent; single-column supplementary pages are detected from long text-line geometry. Each section extends from its detected heading to the next detected top-level heading.
These are approximate layout envelopes, including floats and whitespace—not word counts, occupied-ink areas, exact text-only section lengths or official page-limit accounting. Full-width figures, atypical reading order and page margins can distort the estimate. For that reason an approximately eight-page body envelope need not equal exactly eight. The first numbered section does not include the preceding title, abstract and teaser region.
Main and separate PDFs have independent counts. When the official supplement is a ZIP, its inspected PDF member is identified explicitly; the archive’s video/code content is outside this scan. A missing/uninspected PDF is not assigned zero pages. Figure numbering can restart in supplements, so labels from different files are never collapsed into a single main-paper count.
Accuracy and interpretation
Section titled “Accuracy and interpretation”Caption/heading detection is extraction-assisted. Nonstandard, image-only, shared or continued labels can require source-linked corrections. A numbering gap is a reason to inspect the source, not to invent a missing figure. Every caption and section boundary has not been independently hand-counted. Each paper’s ledger retains locations, extraction notes and any manual correction; the independent sample audit documents its inspected denominator.
Contribution types are editorial classifications based on the paper-specific writing scans, with one primary type per paper. Secondary contributions were not separately coded for this edition; empty secondary lists do not mean papers lack mixed contributions. These are not official CVPR tracks. Selected gallery examples illustrate a useful design; they are not prevalence-ranked visual styles. An accepted-only population cannot establish that any measured layout or writing choice caused acceptance.
Download the structural metadata · Inspect all paper studies · Use a contribution-specific drafting plan.