First of its kind dataset includes more than 300 hours of muscle-signal and first-person video recordings that let robots learn from what hands do beyond what a camera sees

SAN FRANCISCO, Oct. 20, 2026 /PRNewswire/ -- Encord, the data infrastructure company for Physical AI, today released EMG-1, the largest open-source dataset pairing muscle signals to first-person video. Developed in collaboration with Wearable Devices, the release ties biological signals to AI development, using wristbands that read the forearm muscle signals behind every finger movement (a technique called electromyography, or EMG), capturing what the hand is doing even when it is hidden from view. EMG-1 contains more than 300 hours of synchronized recordings during real world tasks alongside optical motion capture; it is over 30 times larger than the next-largest public dataset of its kind.

Autonomous robots learn how to manipulate objects in deployment by training on real-world video of people completing tasks, but cameras lose track of hands when they are occluded, most often by the object being handled. EMG is used to produce accurate hand-pose labels on training data where the camera fails, continually recording the electrical activity that drives each finger. Paired with first-person video, the two streams provide a complete record of a movement that is inaccessible through a camera alone, in addition to signals like force and contact that a camera cannot capture.

Experiments showed that combining EMG with video reduced fingertip tracking error by 25% compared with video alone, and resulted in 35% lower error during short gaps in visual tracking. EMG alone also recognized pinch-like finger movements. Encord is excited to see what the community can do with EMG-1 beyond our initial testing. Researchers can download the dataset and learn more about other initiatives at encord.com/data-labs.

EMG is the first of several biological signals Encord Data Labs is bringing into robot training data. The lab is also conducting research into how brain signals (EEG) can be used to label an operator's mental state, such as satisfaction or error response, and steer robot policies.

"We work with hundreds of companies at the frontier of Physical AI like Toyota, Boston Dynamics and Zipline, and it's clear that real-world model performance comes down to the data going in. With EMG-1 and our new state-of-the-art hand-tracking model, Encord is offering a solution to one of the biggest problems facing robotics teams today," said Eric Landau, Co-Founder and Co-CEO of Encord.

Introducing Encord Data Labs

EMG-1 releases in conjunction with the launch of Encord Data Labs, a research division working at the frontier of Physical AI data development. Encord works with hundreds of leading AI teams, giving the company visibility into their biggest challenges. Data Labs exists to solve them: with innovations in Applied Research & Automation, Advanced Data Capture, and Partnerships, the division makes Physical AI work more reliably in the real world. Alongside EMG-1, Data Labs is also releasing:

Encord-REACH, a new state-of-the-art hand-tracking model built for real-world conditions where existing models break. Tailored for industrial conditions, the model improves tracking with industrial gloves, obscured fingers, fast motion and multiple hands entering the frame. In Encord's evaluations, it reduced tracking errors in wrists by 80%, joints by 65%, and lowered industrial glove errors by up to 70%. It also eliminated mistaken bystander-hand swaps observed in competing models, 27% on average.

Synthetic data generation findings in collaboration with NVIDIA and ANYbotics. The companies generated synthetic examples of rare industrial anomalies, including liquid spills, with NVIDIA Cosmos 3 models to improve anomaly detection in robotic inspection. Using Encord's platform to curate real-world source data and validate synthetic outputs, the collaboration produced training examples for scenarios that are difficult to capture in real-world footage. Synthetic data improved recall on real-world anomalies by up to 24% and increased the examples available for model calibration by 2.2x.

In the coming months, Encord Data Labs will release more findings, including:

- Automated action segmentation for first-person video, which labels task steps from a handful of examples and significantly reduces time spent doing manual labeling.

- A benchmark of frontier vision-language models showing where AI-generated video captions can't yet be trusted for Physical AI.

- Expert-labeled sensor data for Archetype AI's physical world models.

The division's research is fully integrated with Encord's Physical AI data infrastructure, syncing human data with video in Encord so that it's ready to curate, improve and train on. Customers can bring their own teams to generate and enrich training data or work with Encord's experts.

The announcement was made at Encord's inaugural Long Horizon Summit in San Francisco, where Data Labs demonstrated EMG-1 and its hand tracking model live.

"This week, the people building the most ambitious robots in the world saw EMG-1 and our hand tracking running in real time," said Ulrik Stig Hansen, Co-Founder and Co-CEO of Encord. "Data Labs is the home of Encord's research at the frontier of Physical AI. We're just getting started."

About Encord

The company supports over 300 AI teams globally – including Woven by Toyota, Boston Dynamics and Zipline and numerous Physical AI and frontier labs – and has seen significant growth in both revenue and data volume on its platform as a result of the surge in Physical AI.

Media Contact:

Chris Clemens

Nexios Communications Strategies

[email protected]

SOURCE Encord

Share this article