CVPR 2025 Schedule

Skip to yearly menu bar Skip to main content

Filter Events

WED 11 JUN

THU 12 JUN

FRI 13 JUN

9 a.m.

Oral Session 1A: Image and Video Synthesis [9:00-10:15]

Orals 9:00-10:15

[9:00] Motion Prompting: Controlling Video Generation with Motion Trajectories

[9:15] Go-with-the-Flow: Motion-Controllable Video Diffusion Models Using Real-Time Warped Noise

[9:30] LookingGlass: Generative Anamorphoses via Laplacian Pyramid Warping

[9:45] Alias-Free Latent Diffusion Models: Improving Fractional Shift Equivariance of Diffusion Latent Space

[10:00] RandAR: Decoder-only Autoregressive Visual Generation in Random Orders

(ends 10:15 AM)

Oral Session 1B: Interpretability and Evaluation [9:00-10:15]

Orals 9:00-10:15

[9:00] OpenING: A Comprehensive Benchmark for Judging Open-ended Interleaved Image-Text Generation

[9:15] LibraGrad: Balancing Gradient Flow for Universally Better Vision Transformer Attributions

[9:30] Do We Always Need the Simplicity Bias? Looking for Optimal Inductive Biases in the Wild

[9:45] Molmo and PixMo: Open Weights and Open Data for State-of-the-Art Vision-Language Models

[10:00] Rethinking Vision-Language Model in Face Forensics: Multi-Modal Interpretable Forged Face Detector

(ends 10:15 AM)

Oral Session 1C: Image Processing and Deep Architectures [9:00-10:15]

Orals 9:00-10:15

[9:00] CleanDIFT: Diffusion Features without Noise

[9:15] OverLoCK: An Overview-first-Look-Closely-next ConvNet with Context-Mixing Dynamic Kernels

[9:30] Towards Explicit Geometry-Reflectance Collaboration for Generalized LiDAR Segmentation in Adverse Weather

[9:45] DiffFNO: Diffusion Fourier Neural Operator

[10:00] Removing Reflections from RAW Photos

(ends 10:15 AM)

10:30 a.m.

Art Program [10:30-6:00]

(ends 6:00 PM)

1 p.m.

Oral Session 2A: 3D Computer Vision [1:00-2:30]

Orals 1:00-2:30

[1:00] FoundationStereo: Zero-Shot Stereo Matching

[1:15] MoGe: Unlocking Accurate Monocular Geometry Estimation for Open-Domain Images with Optimal Training Supervision

[1:30] Multi-view Reconstruction via SfM-guided Monocular Depth Estimation

[1:45] MV-DUSt3R+: Single-Stage Scene Reconstruction from Sparse Views In 2 Seconds

[2:00] VGGT: Visual Geometry Grounded Transformer

[2:15] CraftsMan3D: High-fidelity Mesh Generation with 3D Native Diffusion and Interactive Geometry Refiner

(ends 2:30 PM)

Oral Session 2B: Human Motion [1:00-2:30]

Orals 1:00-2:30

[1:00] CAP4D: Creating Animatable 4D Portrait Avatars with Morphable Multi-View Diffusion Models

[1:15] Reanimating Images using Neural Representations of Dynamic Stimuli

[1:30] EgoLM: Multi-Modal Language Model of Egocentric Motions

[1:45] Reconstructing Humans with a Biomechanically Accurate Skeleton

[2:00] MEGA: Masked Generative Autoencoder for Human Mesh Recovery

[2:15] TokenHSI: Unified Synthesis of Physical Human-Scene Interactions through Task Tokenization

(ends 2:30 PM)

Oral Session 2C: Temporal Modeling and Action Recognition [1:00-2:30]

Orals 1:00-2:30

[1:00] Descriptor-In-Pixel : Point-Feature Tracking For Pixel Processor Arrays

[1:15] Temporally Consistent Object-Centric Learning by Contrasting Slots

[1:30] Temporal Alignment-Free Video Matching for Few-shot Action Recognition

[1:45] Closed-Loop Supervised Fine-Tuning of Tokenized Traffic Models

[2:00] The PanAf-FGBG Dataset: Understanding the Impact of Backgrounds in Wildlife Behaviour Recognition

[2:15] Rethinking Spiking Self-Attention Mechanism: Implementing α-XNOR Similarity Calculation in Spiking Transformers

(ends 2:30 PM)

5 p.m.

Art Gallery Tour with Curator, Luba Elliott [5:00-5:30]

Chairs: Luba Elliott

(ends 5:30 PM)

SAT 14 JUN

9 a.m.

Oral Session 3A: 3D Computer Vision [9:00-10:15]

Orals 9:00-10:15

[9:00] MegaSaM: Accurate, Fast and Robust Structure and Motion from Casual Dynamic Videos

[9:15] Stereo4D: Learning How Things Move in 3D from Internet Stereo Videos

[9:30] Continuous 3D Perception Model with Persistent State

[9:45] TacoDepth: Towards Efficient Radar-Camera Depth Estimation with One-stage Fusion

[10:00] Neural Inverse Rendering from Propagating Light

(ends 10:15 AM)

Oral Session 3B: Multimodal Computer Vision [9:00-10:15]

Orals 9:00-10:15

[9:00] SegEarth-OV: Towards Training-Free Open-Vocabulary Segmentation for Remote Sensing Images

[9:15] Towards Universal Dataset Distillation via Task-Driven Diffusion

[9:30] IceDiff: High Resolution and High-Quality Arctic Sea Ice Forecasting with Generative Diffusion Prior

[9:45] Efficient Test-time Adaptive Object Detection via Sensitivity-Guided Pruning

[10:00] Keep the Balance: A Parameter-Efficient Symmetrical Framework for RGB+X Semantic Segmentation

(ends 10:15 AM)

Oral Session 3C: Vision and Language [9:00-10:15]

Orals 9:00-10:15

[9:00] Identifying and Mitigating Position Bias of Multi-image Vision-Language Models

[9:15] Mitigating Hallucinations in Large Vision-Language Models via DPO: On-Policy Data Hold the Key

[9:30] Q-Eval-100K: Evaluating Visual Quality and Alignment Level for Text-to-Vision Content

[9:45] Thinking in Space: How Multimodal Large Language Models See, Remember, and Recall Spaces

[10:00] From Multimodal LLMs to Generalist Embodied Agents: Methods and Lessons

(ends 10:15 AM)

11 a.m.

Art Gallery Tour with Curator, Luba Elliott [11:00-11:30]

Chairs: Luba Elliott

(ends 11:30 AM)

1 p.m.

Oral Session 4A: Image and Video Synthesis [1:00-2:15]

Orals 1:00-2:15

[1:00] Reconstruction vs. Generation: Taming Optimization Dilemma in Latent Diffusion Models

[1:15] Language-Guided Image Tokenization for Generation

[1:30] DreamRelation: Bridging Customization and Relation Generation

[1:45] Infinity∞: Scaling Bitwise AutoRegressive Modeling for High-Resolution Image Synthesis

[2:00] Autoregressive Distillation of Diffusion Transformers

(ends 2:15 PM)

Oral Session 4B: Embodied Computer Vision [1:00-2:15]

Orals 1:00-2:15

[1:00] PDFactor: Learning Tri-Perspective View Policy Diffusion Field for Multi-Task Robotic Manipulation

[1:15] RoboSpatial: Teaching Spatial Understanding to 2D and 3D Vision-Language Models for Robotics

[1:30] GROVE: A Generalized Reward for Learning Open-Vocabulary Physical Skill

[1:45] Navigation World Models

[2:00] Viewpoint Rosetta Stone: Unlocking Unpaired Ego-Exo Videos for View-invariant Representation Learning

(ends 2:15 PM)

Oral Session 4C: 3D Computer Vision [1:00-2:15]

Orals 1:00-2:15

[1:00] DORNet: A Degradation Oriented and Regularized Network for Blind Depth Super-Resolution

[1:15] Convex Relaxation for Robust Vanishing Point Estimation in Manhattan World

[1:30] Learned Binocular-Encoding Optics for RGBD Imaging Using Joint Stereo and Focus Cues

[1:45] Camera Resection from Known Line Pencils and a Radially Distorted Scanline

[2:00] Opportunistic Single-Photon Time of Flight

(ends 2:15 PM)

1:30 p.m.

Panel Discussion [1:30-2:30]

(ends 2:30 PM)

5 p.m.

Art Gallery Tour with Curator, Luba Elliott [5:00-5:30]

(ends 5:30 PM)

SUN 15 JUN

9 a.m.

Oral Session 5A: Generative AI [9:00-10:15]

Orals 9:00-10:15

[9:00] Improving Diffusion Inverse Problem Solving with Decoupled Noise Annealing

[9:15] DesignDiffusion: High-Quality Text-to-Design Image Generation with Diffusion Models

[9:30] CustAny: Customizing Anything from A Single Example

[9:45] Minority-Focused Text-to-Image Generation via Prompt Optimization

[10:00] Black-Box Forgery Attacks on Semantic Watermarks for Diffusion Models

(ends 10:15 AM)

Oral Session 5B: Learning Systems and Medical Applications [9:00-10:15]

Orals 9:00-10:15

[9:00] UniAP: Unifying Inter- and Intra-Layer Automatic Parallelism by Mixed Integer Quadratic Programming

[9:15] Geometric Knowledge-Guided Localized Global Distribution Alignment for Federated Learning

[9:30] Enhancing Diversity for Data-free Quantization

[9:45] TopoCellGen: Generating Histopathology Cell Topology with a Diffusion Model

[10:00] Enhancing SAM with Efficient Prompting and Preference Optimization for Semi-supervised Medical Image Segmentation

(ends 10:15 AM)

Oral Session 5C: Visual and Spatial Computing [9:00-10:15]

Orals 9:00-10:15

[9:00] Adv-CPG: A Customized Portrait Generation Framework with Facial Adversarial Attacks

[9:15] Gromov–Wasserstein Problem with Cyclic Symmetry

[9:30] Time of the Flight of the Gaussians: Optimizing Depth Indirectly in Dynamic Radiance Fields

[9:45] Zero-Shot Monocular Scene Flow Estimation in the Wild

[10:00] 3D Student Splatting and Scooping

(ends 10:15 AM)

10:30 a.m.

Art Program [10:30-6:00]

(ends 6:00 PM)

11 a.m.

Art Gallery Tour with Curator, Luba Elliott [11:00-11:30]

Chairs: Luba Elliott

(ends 11:30 AM)

1 p.m.

Oral Session 6A: 3D from Single or Multi-View Sensors [1:00-2:30]

Orals 1:00-2:15

[1:00] DIFIX3D+: Improving 3D Reconstructions with Single-Step Diffusion Models

[1:15] 3DGUT: Enabling Distorted Cameras and Secondary Rays in Gaussian Splatting

[1:30] DNF: Unconditional 4D Generation with Dictionary-based Neural Fields

[1:45] CAT4D: Create Anything in 4D with Multi-View Video Diffusion Models

[2:00] Diffusion Renderer: Neural Inverse and Forward Rendering with Video Diffusion Models

(ends 2:30 PM)

Oral Session 6B: Scene Understanding, Image Editing and Multimodal Learning [1:00-2:30]

Orals 1:00-2:30

[1:00] Effective SAM Combination for Open-Vocabulary Semantic Segmentation

[1:15] FluidNexus: 3D Fluid Reconstruction and Prediction from a Single Video

[1:30] Birth and Death of a Rose

[1:45] Semi-Supervised State-Space Model with Dynamic Stacking Filter for Real-World Video Deraining

[2:00] AnyEdit: Mastering Unified High-Quality Image Editing for Any Idea

[2:15] Generative Multimodal Pretraining with Discrete Diffusion Timestep Tokens

(ends 2:30 PM)

Oral Session 6C: Video, Action, and Language [1:00-2:30]

Orals 1:00-2:30

[1:00] Seeing Far and Clearly: Mitigating Hallucinations in MLLMs with Attention Causal Decoding

[1:15] Video-XL: Extra-Long Vision Language Model for Hour-Scale Video Understanding

[1:30] LoRASculpt: Sculpting LoRA for Harmonizing General and Specialized Knowledge in Multimodal Large Language Models

[1:45] VideoEspresso: A Large-Scale Chain-of-Thought Dataset for Fine-Grained Video Reasoning via Core Frame Selection

[2:00] SEAL: Semantic Attention Learning for Long Video Representation

[2:15] Learning Audio-guided Video Representation with Gated Attention for Video-Text Retrieval

(ends 2:30 PM)

1:30 p.m.

Panel Discussion [1:30-2:30]

(ends 2:30 PM)