Keep It in Mind: User Centric Continual Spatial Intelligence Reasoning in Egocentric Video Streams
Paper • 2606.15200 • Published
UCS-Bench is a benchmark for evaluating user-centric continual spatial intelligence in streaming egocentric videos.
The dataset contains 170+ hours of egocentric visual observations and 8.1K+ timestamped questions. It is designed to test whether models can understand dynamic spatial environments from a user's point of view, maintain long-term spatial memory, and answer questions grounded in the user's real-time movement and location.
Unlike standard video QA benchmarks, UCS-Bench focuses on long-horizon spatial reasoning in continuous egocentric streams, including:
Each video is paired with metadata in metadata.jsonl.
Example:
{"file_name": "video001.mp4", "label": "xxx"}