Research at the Frontier

[Applied Research & Automation]
[Advanced Data Capture]
[Partnerships]

[ LATEST LAUNCH ]

EMG-1

The largest open-source dataset of muscle signals synced to egocentric video. EMG-1 delivers major improvements to hand tracking — teaching robotic hands to perform complex tasks by solving occlusion issues and adding force and contact data invisible to cameras.

Read the article

A hand wearing an EMG sensor band, overlaid with a dot grid

Pushing the boundaries of Physical AI data

[01]Applied Research & AutomationDeveloping models and automation to interpret complex Physical AI data at scale.

[02]Advanced Data CaptureCapturing and synchronizing new sensor modalities like EMG to complement vision.

[03]PartnershipsWorking with customers and industry leaders to test new approaches and deliver real-world results.

Research articles

[Applied Research & Automation]

Introducing Encord Data Labs

A research division advancing how physical-world data is captured, understood and evaluated, starting with hand tracking and new sensing modalities.

[Applied Research & Automation]

Hands at Work: Egocentric hand pose tracking for real world scenes

Hand tracking for humanoid robotics has to survive bystanders, thick safety gloves and multimodal capture rigs. Here is how Encord-REACH approaches each.

[Advanced Data Capture]

Introducing EMG-1: the largest open dataset of wrist EMG synchronized with egocentric video

EMG-1 pairs 307 hours of field wrist EMG with synchronized egocentric video, alongside a 10.4-hour laboratory collection with optical motion capture.

Subscribe for new releases

Research at the frontier.