mlsanityData to 14 Oct 2026

EgoCross: Benchmarking Multimodal Large Language Models for Cross-Domain Egocentric Video Question Answering

arXiv 2508.10729 ↗First posted 14 Aug 2025cs.CV cs.AIcited by 21

On the map: Multimodal LLMs / vision-language models → Embodied and 3D vision-language models

Abstract (arXiv)

Recent advances in Multimodal Large Language Models (MLLMs) have significantly pushed the frontier of egocentric video question answering (EgocentricQA). However, existing benchmarks and studies are mainly limited to common daily activities such as cooking and cleaning. In contrast, real-world deployment inevitably encounters domain shifts, where target domains differ substantially in both visual style and semantic content. To bridge this gap, we introduce EgoCross, a comprehensive benchmark designed to evaluate the cross-domain generalization of MLLMs in EgocentricQA. EgoCross covers four diverse and challenging domains, including surgery, industry, extreme sports, and animal perspective, representing realistic and high-impact application scenarios. It comprises approximately 1,000 QA pairs across 798 video clips, spanning four key QA tasks: prediction, recognition, localization, and counting. Each QA pair provides both OpenQA and CloseQA formats to support fine-grained evaluation. Extensive experiments show that most existing MLLMs, whether general-purpose or egocentric-specialized, struggle to generalize to domains beyond daily life, highlighting the limitations of current models. Furthermore, we conduct several pilot studies, e.g., fine-tuning and reinforcement learning, to explore potential improvements. We hope EgoCross and our accompanying analysis will serve as a foundation for advancing domain-adaptive, robust egocentric video understanding.

Key facts. Cited by 21 papers in the map (data through 14 Oct 2026). Placed in Embodied and 3D vision-language models, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 40 of its 51 references that are in the map. First posted on arXiv 14 Aug 2025.

Builds on

40 of 51 references.

Closest ancestors · 2

Same direction, newest first

  1. 2025Omnia de EgoTempo: Benchmarking Temporal Understanding of Multi-Modal LLMs in Egocentric Videos17 Mar
  2. 2023EgoThink: Evaluating First-Person Perspective Thinking Capability of Vision-Language Models27 Nov

Other references · 38

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
  2. 2023GPT-4 Technical Reportcited by 18,396
  3. 2023Efficient Memory Management for Large Language Model Serving with PagedAttentioncited by 5,944
  4. 2025Qwen2.5-VL Technical Reportcited by 5,587
  5. 2025Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilitiescited by 3,796
  6. 2025InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Modelscited by 1,834
  7. 2023EgoSchema: A Diagnostic Benchmark for Very Long-form Video Language Understandingcited by 680
  8. 2025VideoLLaMA 3: Frontier Multimodal Foundation Models for Image and Video Understandingcited by 518
  9. 2023EgoVLPv2: Egocentric Video-Language Pre-training with Fusion in the Backbonecited by 130
  10. 2022EgoTaskQA: Understanding Human Tasks in Egocentric Videoscited by 115
  11. 2025EgoLife: Towards Egocentric Life Assistantcited by 91
  12. 2022Video Question Answering: Datasets, Algorithms and Challengescited by 84
26 more
  1. 2020Cross-domain Object Detection through Coarse-to-Fine Feature Adaptationcited by 82
  2. 2023StyleAdv: Meta Style Adversarial Training for Cross-Domain Few-Shot Learningcited by 53
  3. 2019Adversarial Cross-Domain Action Recognition with Co-Attentioncited by 51
  4. 2021Meta-FDMixup: Cross-Domain Few-Shot Learning Guided by Labeled Target Datacited by 48
  5. 2025EgoTextVQA: Towards Egocentric Scene-Text Aware Video Question Answeringcited by 44
  6. 2024Cross-Domain Few-Shot Object Detection via Enhanced Open-Set Object Detectorcited by 43
  7. 2022Aligning Domain-specific Distribution and Classifier for Cross-domain Classification from Multiple Sourcescited by 42
  8. 2023M3Net: Multi-view Encoding, Matching, and Fusion for Few-shot Fine-grained Action Recognitioncited by 34
  9. 2021Aligning Correlation Information for Domain Adaptation in Action Recognitioncited by 30
  10. 2022Scene-aware Egocentric 3D Human Pose Estimationcited by 30
  11. 2025NTIRE 2025 Challenge on Cross-Domain Few-Shot Object Detection: Methods and Resultscited by 25
  12. 2022Wave-SAN: Wavelet based Style Augmentation Network for Cross-Domain Few-Shot Learningcited by 23
  13. 2024EgoSurgery-Phase: A Dataset of Surgical Phase Recognition from Egocentric Open Surgery Videoscited by 20
  14. 2022Locate before Answering: Answer Guided Question Localization for Video Question Answeringcited by 19
  15. 2023From Global to Local: Multi-scale Out-of-distribution Detectioncited by 18
  16. 2023CholecTrack20: A Multi-Perspective Tracking Dataset for Surgical Toolscited by 17
  17. 2025Reliable Few-shot Learning under Dual Noisescited by 17
  18. 2022TGDM: Target Guided Dynamic Mixup for Cross-Domain Few-Shot Learningcited by 16
  19. 2025Domain-RAG: Retrieval-Guided Compositional Image Generation for Cross-Domain Few-Shot Object Detectioncited by 15
  20. 2024Multimodal Cross-Domain Few-Shot Learning for Egocentric Action Recognitioncited by 14
  21. 2025CLiViS: Unleashing Cognitive Map through Linguistic-Visual Synergy for Embodied Visual Reasoningcited by 13
  22. 2024EgoPet: Egomotion and Interaction Data from an Animal's Perspectivecited by 11
  23. 2024Prompt as Free Lunch: Enhancing Diversity in Source-Free Cross-domain Few-shot Learning through Semantic-Guided Promptingcited by 10
  24. 2024Human-Centric Transformer for Domain Adaptive Action Recognitioncited by 8
  25. 2016Trajectory Aligned Features For First Person Action Recognitioncited by 7
  26. 2025Cross-View Multi-Modal Segmentation @ Ego-Exo4D Challenges 2025cited by 5

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 4; 2026: 17.

21 papers.

Sort:

By area:

  1. 2025Multimodal Spatial Reasoning in the Large Model Era: A Survey and Benchmarkscited by 33
  2. 2025EgoNight: Towards Egocentric Vision Understanding at Night with a Challenging Benchmarkcited by 19
  3. 2025StreamEQA: Towards Streaming Video Understanding for Embodied Scenarioscited by 12
  4. 2026EgoSound: Benchmarking Sound Understanding in Egocentric Videoscited by 11
  5. 2025MICA: Multi-Agent Industrial Coordination Assistantcited by 8
  6. 2026VisNec: Measuring and Leveraging Visual Necessity for Multimodal Instruction Tuningcited by 8
  7. 2026PanoEnv: Exploring 3D Spatial Intelligence in Panoramic Environments with Reinforcement Learningcited by 7
  8. 2026OmniEgo-R2: A Routed Reasoning Framework for the 1st Cross-Domain EgoCross Challenge at CVPR 2026cited by 5
  9. 2026IMPACT-CYCLE: A Contract-Based Multi-Agent System for Claim-Level Supervisory Correction of Long-Video Semantic Memorycited by 2
  10. 2026EXPLORE-Bench: Egocentric Scene Prediction with Long-Horizon Reasoningcited by 1
  11. 2026EgoEsportsQA: An Egocentric Video Benchmark for Perception and Reasoning in Esportscited by 1
  12. 2026Once-For-All: A Train-Once and Select-Anytime Framework for Multimodal Instruction Tuningcited by 1
9 more
  1. 2026The Right Inference Strategy Is All You Need: Nearly Training-Free Domain-Wise Inference for EgoCross Challengecited by 1
  2. 2026MEMO: Multi-Level Entity-Aware Memory for Streaming Video Understandingcited by 1
  3. 2026IMPACT-Scribe: Interactive Temporal Action Segmentation with Boundary Scribbles and Query Planningcited by 0
  4. 2026Reflective Dialogue between Teacher and Solver Agents for Video Question Answeringcited by 0
  5. 2026Reinforcing Egocentric Spatial Perception in Multimodal Large Language Models via Ego Scene Augmentationcited by 0
  6. 2026The First EgoCross Challenge at EgoVis 2026: Cross-Domain Egocentric Video Question Answeringcited by 0
  7. 2026Does Local Video Understanding Transfer Across Encounters? The EgoGears Benchmarkcited by 0
  8. 2026EgoProbe: Benchmarking Long-Range and Exact Memory in Streaming Egocentric Videoscited by 0
  9. 2026EgoOmniWorld: Unifying Egocentric Understanding via Ego-Action-World Interactionscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Advancing Egocentric Video Question Answering with Multimodal Large Language Models
  2. 2024MM-Ego: Towards Building Egocentric Multimodal LLMs for Video QA
  3. 2026Graph it first! Enabling Reasoning on Long-form Egocentric Videos through Scene Graphs
  4. 2023LifelongMemory: Leveraging LLMs for Answering Queries in Long-form Egocentric Videos
  5. 2026Ego-Grounding for Personalized Question-Answering in Egocentric Videos
  6. 2026MA-EgoQA: Question Answering over Egocentric Videos from Multiple Embodied Agents
  7. 2025How Far Can Off-the-Shelf Multimodal Large Language Models Go in Online Episodic Memory Question Answering?
  8. 2024LongVideoBench: A Benchmark for Long-context Interleaved Video-Language Understanding