CVPR 2025

2,907 papers

Top Research Directions

Research directions most represented by CVPR 2025 papers.

Overcoming the Static Scene Assumption 49
Mitigating Multi View Inconsistency in Generative 3D Modeling 48
Eliminating Per-Scene Optimization with Feed-Forward 3D Reconstruction 42
Generalizing from Synthetic Data to the Real World 39
Material Lighting Disentanglement for 3D Reconstruction 37
Spatial Attention Masking in Generative Diffusion Models 36
Error Propagation in Monocular Depth Priors 32
Overcoming the Iterative Inference Latency of Diffusion Models 31
Exploiting Temporal Redundancy in Video Neural Networks 29
Aligning Global Semantics with Local Spatial Precision 28
Scaling Visual Attention Beyond Quadratic Complexity 28
Accelerating Neural Scene Representations for Real Time Rendering 27
Visual Grounding and Hallucination in Vision Language Models 27
Completing Unobserved Geometry with Generative Diffusion Priors 24
The Cross Modal Representation Gap 24

Top Authors

AuthorPapers
Yujun Shen17
Marc Pollefeys15
Daniel Cremers13
Jian Yang13
Jan Kautz11
Ziwei Liu11
Federico Tombari10
Guangtao Zhai10
Jiangning Zhang10
Christian Theobalt9
Jiajun Wu9
Ming-Hsuan Yang9
Ranjay Krishna9
Wei-Shi Zheng9
Xiongkuo Min9
Zhe Lin9
Zheng-Jun Zha9
Andrea Vedaldi8
Chen Change Loy8
Cheng Wang8

Impactful Papers

Papers from CVPR 2025 that shaped research clusters.

Foundational

Influential

Papers

Title
TopV: Compatible Token Pruning with Inference Time Optimization for Fast and Low-Memory Multimodal Vision Language Model
Tora: Trajectory-oriented Diffusion Transformer for Video Generation
Touch2Shape: Touch-Conditioned 3D Diffusion for Shape Exploration and Reconstruction
Toward Generalized Image Quality Assessment: Relaxing the Perfect Reference Quality Assumption
Toward Real-world BEV Perception: Depth Uncertainty Estimation via Gaussian Splatting
Toward Robust Neural Reconstruction from Sparse Point Sets
Towards All-in-One Medical Image Re-Identification
Towards a Universal Synthetic Video Detector: From Face or Background Manipulations to Fully AI-Generated Content
Towards Better Alignment: Training Diffusion Models with Reinforcement Learning Against Sparse Rewards
Towards Consistent Multi-Task Learning: Unlocking the Potential of Task-Specific Parameters
Towards Continual Universal Segmentation
Towards Cost-Effective Learning: A Synergy of Semi-Supervised and Active Learning
Towards Effective and Sparse Adversarial Attack on Spiking Neural Networks via Breaking Invisible Surrogate Gradients
Towards Efficient Foundation Model for Zero-shot Amodal Segmentation
Towards Enhanced Image Inpainting: Mitigating Unwanted Object Insertion and Preserving Color Consistency
Towards Explainable and Unprecedented Accuracy in Matching Challenging Finger Crease Patterns
Towards Explicit Geometry-Reflectance Collaboration for Generalized LiDAR Segmentation in Adverse Weather
Towards Fine-Grained Interpretability: Counterfactual Explanations for Misclassification with Saliency Partition
Towards Generalizable Scene Change Detection
Towards Generalizable Trajectory Prediction using Dual-Level Representation Learning and Adaptive Prompting
Towards General Visual-Linguistic Face Forgery Detection
Towards High-fidelity 3D Talking Avatar with Personalized Dynamic Texture
Towards Human-Understandable Multi-Dimensional Concept Discovery
Towards Improved Text-Aligned Codebook Learning: Multi-Hierarchical Codebook-Text Alignment with Long Text
Towards In-the-wild 3D Plane Reconstruction from a Single Image
Towards Long-Horizon Vision-Language Navigation: Platform, Benchmark and Method
Towards Lossless Implicit Neural Representation via Bit Plane Decomposition
Towards Million-Scale Adversarial Robustness Evaluation With Stronger Individual Attacks
Towards More General Video-based Deepfake Detection through Facial Component Guided Adaptation for Foundation Model
Towards Natural Language-Based Document Image Retrieval: New Dataset and Benchmark
Towards Open-Vocabulary Audio-Visual Event Localization
Towards Optimizing Large-Scale Multi-Graph Matching in Bioimaging
Towards Practical Real-Time Neural Video Compression
Towards Precise Embodied Dialogue Localization via Causality Guided Diffusion
Towards Precise Scaling Laws for Video Diffusion Transformers
Towards RAW Object Detection in Diverse Conditions
Towards Realistic Example-based Modeling via 3D Gaussian Stitching
Towards Satellite Image Road Graph Extraction: A Global-Scale Dataset and A Novel Method
Towards Scalable Human-aligned Benchmark for Text-guided Image Editing
Towards Smart Point-and-Shoot Photography
Towards Source-Free Machine Unlearning
Towards Stable and Storage-efficient Dataset Distillation: Matching Convexified Trajectory
Towards Training-free Anomaly Detection with Vision and Language Foundation Models
Towards Transformer-Based Aligned Generation with Self-Coherence Guidance
Towards Unbiased and Robust Spatio-Temporal Scene Graph Generation and Anticipation
Towards Understanding and Quantifying Uncertainty for Text-to-Image Generation
Towards Understanding How Knowledge Evolves in Large Vision-Language Models
Towards Universal AI-Generated Image Detection by Variational Information Bottleneck Network
Towards Universal Dataset Distillation via Task-Driven Diffusion
Towards Universal Fake Image Detectors that Generalize Across Generative Models

« prev Page 53/59 next »