EgoExplorer addresses a core challenge in humanoid learning: capturing how people coordinate whole-body movement and fine hand manipulation during real-world tasks. Its wearable system combines first-person video, body tracking and glove measurements to record continuous human demonstrations across laboratory, household and industrial settings. MANUS Metagloves Pro provide detailed finger articulation within this capture pipeline.
The researchers integrated MANUS Metagloves Pro alongside five VIVE Ultimate Trackers and a head-mounted camera. Participants wore the system while performing tasks in their working environments, allowing the team to record finger movements together with the posture and locomotion that accompany manipulation.
The gloves directly measured finger articulation, providing hand-motion data beyond what could be observed in thecamera view. EgoExplorer mapped the glove skeletons onto the hands of an SMPL-X digital human model, aligning the recorded finger movements with the reconstructed body. The EgoExplorer research team described MANUS Metagloves Pro as “the backbone of our hand stream.”
The pipeline synchronized hand motion, reconstructed whole-body movement and first-person video on a shared 30 Hz timeline. The resulting dataset supports research into the relationship between visual observations and human actions.
EgoExplorer illustrates how MANUS hand-motion capture can contribute to embodied AI research by recording detailed finger movements as part of continuous human demonstrations, connecting hand manipulation with whole-body activity and visual context.
