Ego4D: Around the World in 3,000 Hours of Egocentric Video
PulseAugur coverage of Ego4D: Around the World in 3,000 Hours of Egocentric Video — every cluster mentioning Ego4D: Around the World in 3,000 Hours of Egocentric Video across labs, papers, and developer communities, ranked by signal.
3 day(s) with sentiment data
-
Future context improves gaze estimation, but only up to a point
Researchers have developed a framework to study the impact of future video frames on egocentric gaze estimation models. Their findings indicate that while future context improves causal gaze prediction, the benefits pla…
-
New dataset and model advance scene graph reasoning for human activity understanding
Researchers have introduced SG-Ego, a new dataset that extends Ego4D with spatio-temporal scene graphs to better understand human activities in first-person videos. They also developed GLEN, a graph-based model designed…
-
New benchmark LongEgoRefer challenges AI with long-form egocentric video comprehension
Researchers have introduced LongEgoRefer, a new benchmark designed to evaluate video referring expression comprehension in long-form egocentric videos. This benchmark, derived from the Ego4D dataset, features nearly 1,5…
-
FlexLAM introduces variable-length latent actions to improve video-based decision-making
Researchers have introduced FlexLAM, a novel approach to latent action learning that addresses the bottleneck trade-off in existing models. Unlike previous methods that use a fixed-capacity bottleneck, FlexLAM employs v…
-
New benchmark and architectures for proactive AI assistants released
Researchers have introduced EgoProactive, a new dataset and benchmark suite called Pro extsuperscript{2}Bench, designed to evaluate proactive procedural assistance systems. These systems aim to provide real-time, step-b…
-
New method fuses hand trajectory for egocentric video query grounding
Researchers have developed a new method for grounding natural language queries in egocentric videos by incorporating hand trajectory data. This approach fuses hand kinematic features with pre-trained video-text features…
-
FROST-STA system predicts object interactions in egocentric video
Researchers have developed FROST-STA, a system designed for short-term anticipation in egocentric videos, aiming to predict object interactions. The model uses frozen dense features from a ViT-G backbone, extracting vid…
-
New model advances behavioral recognition from AR glasses sensors
Researchers have developed a new method for recognizing complex human behaviors using data from head-mounted Inertial Measurement Units (IMUs), commonly found in AR smart glasses. They created a large dataset and a hier…
-
VISTA system wins Ego4D challenge with object interaction anticipation
Researchers have developed VISTA, a novel system designed for anticipating human-object interactions in egocentric videos. VISTA integrates spatial object detection with temporal context from a frozen V-JEPA 2.1 model t…