云创展汇

ENGLISH
ENGLISH
DEUTSCH
日本語
|
Sign In
Start

Solutions

About Us

Services

News Center  >

News Detail

Meta has updated its research on sEMG wristband for input, published in Nature.

Source:Internet
2025-07-25 05:44:57

The latest issue of Nature published Meta’s newly peer-reviewed article, which details the company’s R&D in this area and validates sEMG as an intuitive and seamless input method applicable to most people.



Meta believes that sEMG not only enables intuitive and seamless mobile interaction with devices but also supports the work of external research laboratories. These studies indicate that the technology is inherently inclusive, as it is suitable for people with diverse physical abilities and characteristics.


Meta has successfully developed an sEMG wristband prototype in conjunction with Orion—its first true AR glasses—and this is only the beginning. The research team has developed advanced machine learning models capable of converting neural signals that control wrist muscles into commands that drive user interaction with the glasses, thereby eliminating reliance on traditional and cumbersome input methods. This allows users to type and send messages without a keyboard, navigate menus without a mouse, and even clearly observe their surroundings while interacting with digital content—no longer needing to look down at their phones.



sEMG can recognize the user's intent to perform a variety of gestures, such as tapping, swiping, and pinching—all of which can be done with the hand naturally hanging at the side. With Meta's handwriting recognition technology, users can also quickly record information on hard surfaces like desktops, tables, or even legs, opening up new possibilities for discreet communication on the move.


Meta's neural network is trained on data from thousands of volunteer study participants, enabling it to accurately decode subtle gestures across a wide range of people without the need for individual calibration. While the general model already delivers excellent out-of-the-box performance, minimal personalized adjustments based on limited personal data can improve handwriting recognition accuracy by up to 16%—in other words, the sEMG wristband can adapt to users over time and continuously optimize its performance.


It is completely non-invasive, pioneering a new way to interact with computers using muscle signals while addressing many of the problems faced by other forms of human-computer interaction (HCI).


It is convenient, simple, and natural to use—and in some cases, more practical or appropriate than alternative methods like voice interaction, such as when sending private messages in public.


It is always available and eliminates the need for bulky accessories that can take you out of the moment and distract you from the most important people and things.


Meta stated that the paper published in Nature provides a blueprint for the broader scientific community to create their own neuromotor interfaces. In addition to outlining a set of key design rules and best practices covering hardware, experimental design, data requirements, and modeling, Meta has also publicly released a dataset containing over 100 hours of sEMG recordings from more than 300 study participants across three different tasks. Combined with Meta's previously open-sourced sEMG datasets for pose estimation and surface input, they hope today's release will help accelerate future work in the field by academics and researchers.

Other News

  • Capture Cloud Hyper-Realistic Digital Humans Usher in a New Era of AIGC and Redefine Virtual Interactive Experiences

  • The 3D scanning tool "RealityCapture" has been renamed "RealityScan", and its new version will feature updated AI functions

  • A New Era of "Cloud Factory Tours"! Puleisi × Capture Cloud: Reconstructing Factory Visit Models with 3D Real Scenes