EgoActionBench: Benchmarking Dense Action Captioning for Egocentric Human Videos
cs.CV
Submitted: 2026-11-02
Updated: 2026-11-02
Code: https://github.com/player0718/EgoActionBench
Project page: https://egoaction-bench.github.io
Terminology
Sources
- LLaVA-OneVision-2: Towards Next-Generation Perceptual Intelligence
- Qwen3-VL Technical Report
- Qwen2.5-VL Technical Report
- ACE-Data-0: Human-Centric Ambient Capture as Embodied Data Engine
- Molmo2: Open Weights and Data for Vision-Language Models with Video Understanding and Grounding
- RynnBrain: Open Embodied Foundation Models
- HumanNet: Scaling Human-centric Video Learning to One Million Hours
- Cooking beyond Frames: A Stereo Event Camera Dataset in the Kitchen
- DreamDojo: A Generalist Robot World Model from Large-Scale Human Videos
- The "something something" video database for learning and evaluating visual common sense
- Video ReCap: Recursive Captioning of Hour-Long Videos
- The Kinetics Human Action Video Dataset
- FLORA: An Open Model Suite for Scaling Robotics Data Engines
- EgoLive: A Large-Scale Egocentric Dataset from Real-World Human Tasks
- Open-AoE: An Open Egocentric Manipulation Dataset and Toolchain for Embodied Learning
- CAPEval: A Decoupled Caption Evaluation across Understanding and Generation
- Understanding R1-Zero-Like Training: A Critical Perspective
- Ego-OSCAR: Egocentric Open source Stereo CAptuRe System
- Modeling Fine-Grained Hand-Object Dynamics for Egocentric Video Representation Learning
- EgoThinker: Unveiling Egocentric Reasoning with Spatio-Temporal CoT
Related papers
- TEA: Text Encoder Alignment for Robust Concept Erasure in Text-to-Image Models
- Do Egocentric Video-Language Models Capture Both Hand- and Object-Centric Cues?
- Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment
- PVRF: All-in-one Adverse Weather Removal via Prior-modulated and Velocity-constrained Rectified Flow
- DLWM: Diverse Latent World Models for Efficient Multimodal Reasoning
- Out of Sight, Still in Mind: Token Compression for Omni-LLMs