The first EgoCross Challenge, held at EgoVis 2026 during CVPR 2026, introduced a benchmark for evaluating multimodal large language models on cross-domain egocentric video question answering. The challenge focused on models' ability to generalize to diverse domains such as surgery, industrial assembly, extreme sports, and animal perspectives. Participants submitted over 1,500 entries across two tracks: Source-Limited and Open-Source, with the goal of advancing egocentric video understanding. AI
IMPACT Establishes a new benchmark for multimodal LLMs in egocentric video understanding, potentially driving advancements in generalization capabilities.
RANK_REASON Academic paper introducing a new benchmark and challenge. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
