ECCV 2026 accepted papers on Vision-language and multimodal
583 of the 2,834 papers accepted at ECCV 2026 (20.6%) are on Vision-language and multimodal, compared with 10.2% at ECCV 2024. They fall into 6 topics.
See ECCV 2026 in the interactive map
Working on something in this area? Describe your idea in scime atlas to see the closest papers by meaning and how crowded the spot has become.
Topics and papers
Multimodal LLMs · 334 papers
- ViewSpatial-Bench: Evaluating Multi-perspective Spatial Localization in Vision-Language Models (128 citations)
- Video-Holmes: Can MLLM Think like Holmes for Complex Video Reasoning? (111 citations)
- Visual Spatial Tuning (103 citations)
- Chain-of-Visual-Thought: Teaching VLMs to See and Think Better with Continuous Visual Tokens (80 citations)
- LEGO-Puzzles: How Good Are MLLMs at Multi-Step Spatial Reasoning? (53 citations)
- Vision-as-Inverse-Graphics Agent via Interleaved Multimodal Reasoning (53 citations)
- TIR-Bench: A Comprehensive Benchmark for Agentic Thinking-with-Images Reasoning (34 citations)
- SRUM: Fine-Grained Self-Rewarding for Unified Multimodal Models (32 citations)
- BBQ-V: Benchmarking Visual Stereotype Bias in Large Multimodal Models (31 citations)
- UniREditBench: A Unified Reasoning-based Image Editing Benchmark (31 citations)
- Embed-RL: Reinforcement Learning for Reasoning-Driven Multimodal Embeddings (28 citations)
- Towards Spatial Trace with Reasoning in Vision-Language Models for Robotics (22 citations)
and 322 more in the interactive map.
Robot action and navigation · 93 papers
- VLA-JEPA: Enhancing Vision-Language-Action Model with Latent World Model (131 citations)
- RoboClaw: An Agentic Framework for Scalable Long-Horizon Robotic Tasks (57 citations)
- VLA-R1: Enhancing Reasoning in Vision-Language-Action Models (50 citations)
- EchoVLA: Robotic Vision-Language-Action Model with Synergistic Declarative Memory for Mobile Manipulation (34 citations)
- Robobench: A Comprehensive Evaluation Benchmark for Multimodal Large Language Models as Embodied Brain (30 citations)
- FindingDory: A Benchmark to Evaluate Memory in Embodied Agents (26 citations)
- MindDrive: A Vision-Language-Action Model for Autonomous Driving via Online Reinforcement Learning (26 citations)
- UniDrive-WM: Unified Understanding, Planning and Generation World Model For Autonomous Driving (26 citations)
- Scaling Verification Can Be More Effective than Scaling Policy Learning for Vision-Language-Action Alignment (25 citations)
- Towards Generalizable Robotic Manipulation in Dynamic Environments (21 citations)
- MobileVLA-R1: Reinforcing Vision-Language-Action for Mobile Robots (16 citations)
- LIBERO-Safety: A Comprehensive Benchmark for Physical and Semantic Safety in Vision-Language-Action Models (15 citations)
and 81 more in the interactive map.
Cross-modal learning · 67 papers
- ChronusOmni: Improving Time Awareness of Omni-Modal Large Language Models (10 citations)
- UniDriveDreamer: A Single-Stage Multimodal World Model for Autonomous Driving (10 citations)
- ELVA: Exploring Ranking-Driven Universal Multimodal Retrieval (7 citations)
- Histopathology Multi-modal Embedding for Pathology Composed Retrieval (5 citations)
- Together, Then Apart: Balancing Alignment and Distinctiveness for Multimodal Survival Analysis (5 citations)
- Delayed Bidirectional Alignment via Disentangled Audio Semantics for Audio-Visual Segmentation (4 citations)
- Don't Let the Video Speak: Audio-Contrastive Preference Optimization for Audio-visual Language Models (4 citations)
- UniStitch: Unifying Semantic and Geometric Features for Image Stitching (3 citations)
- Divide and Conquer: Decoupled Representation Alignment for Multimodal World Models (2 citations)
- General Incomplete Multimodal Learning via Dynamic Quality Perception (2 citations)
- LaVPR: Benchmarking Language and Vision for Place Recognition (2 citations)
- MJEPA: A Simple and Scalable Joint-Embedding Predictive Architecture for Audio-Visual Learning (2 citations)
and 55 more in the interactive map.
CLIP and zero-shot · 57 papers
- Comprehensive language–image pre-training for 3D medical image understanding (28 citations)
- Personalize Your Large Vision-language Models With In-context Prompt Tuning (10 citations)
- AutoV: Loss-Oriented Ranking for Visual Prompt Retrieval in LVLMs (7 citations)
- OmniFall: From Staged Through Synthetic to Wild, A Unified Multi-Domain Dataset for Robust Fall Detection (7 citations)
- MMLoP: Multi-Modal Low-Rank Prompting for Efficient Vision-Language Adaptation (6 citations)
- VLOD-TTA: Test-Time Adaptation of Vision-Language Object Detectors (5 citations)
- OpenGround: Planning-based Online Perception for Open-World 3D Visual Grounding (4 citations)
- ARGENT: Adaptive Hierarchical Image-Text Representations (3 citations)
- Defect-aware Hybrid Prompt Optimization for Zero-Shot Multi-type Anomaly Detection and Segmentation (3 citations)
- What CLIP Knows but Cannot Say: Recovering Negation from Frozen Intermediate Features (3 citations)
- DeltaDeno: Zero-Shot Anomaly Generation via Delta-Denoising Attribution (2 citations)
- DroneFINE: Domain-Aware Parameter-Efficient Fine-Tuning of Vision-Language Detectors for Drone Images (2 citations)
and 45 more in the interactive map.
Human and embodied modeling · 23 papers
- UniRec-0.1B: Unified Text and Formula Recognition with 0.1B Parameters (14 citations)
- CFM: Language-aligned Concept Foundation Model for Vision (5 citations)
- Cheers: Decoupling Patch Details from Semantic Representations Enables Unified Multimodal Comprehension and Generation (5 citations)
- RGBT-GroundBench: Visual Grounding Beyond RGB in Complex Real-World Scenarios (5 citations)
- UniMotion: A Unified Framework for Motion-Text-Vision Understanding and Generation (3 citations)
- ViQ: Text-Aligned Visual Quantized Representations at Any Resolution (3 citations)
- Discrete Diffusion Models with MLLMs for Unified Medical Multimodal Generation (2 citations)
- FlowInOne: Unifying Multimodal Generation as Image-in, Image-out Flow Matching (2 citations)
- Let ViT Speak: Generative Language-Image Pre-training (2 citations)
- UniDDT: Unifying Multimodal Understanding and Generation with Decoupled Diffusion Transformer (2 citations)
- Evaluating the Interpretability of Sparse Autoencoders with Concept Annotations (1 citation)
- Learning Structured Visual Compositional Representations for Weakly Supervised Referring Expression Comprehension (1 citation)
and 11 more in the interactive map.
Visual question answering · 9 papers
- ProMSA:Progressive Multimodal Search Agents for Knowledge-Based Visual Question Answering (19 citations)
- Diagram-MMU: A Multi-Modal Benchmark for Scientific Diagrams (3 citations)
- Identifying and Resolving Pitfalls of Knowledge-Based VQA Benchmarks: Auditing, Repairing, and Augmenting (1 citation)
- What You Ask is What You Ground: Bridging Question Intent to Temporal Evidence for Grounded VideoQA (1 citation)
- Attention-Logit Steering to Compositional Generalization for Continual VQA
- CRAG-MM-Diagnostics: Enabling Stage-Wise Analysis of Knowledge-Intensive VQA
- Q-TriM: Question-Guided Tri-Modal Attention for Audio–Visual Question Answering
- SAFE-EQA: Semantic-Aware Efficient Exploration for Embodied Question Answering
- MMAgent-R2: Learning to Rerank and Reject for Agentic mRAG
