Meta’s Reality Labs has published a peer-reviewed article that describes the company’s efforts to build an input tool that is non-invasive and relies on surface electromyography (sEMG). The company has been working on wristband technology that is designed to enable human-computer interaction (HCI) using muscle signals. Meta says that the technology allows the wristband to capture a user’s “intent” to pinch, tap, or swipe, using their fingers, or type and navigate on a computer without using a keyboard or a mouse.
A peer-reviewed article in the journal Nature describes the experimental technology being used by the company, that relies on a wristband designed to recognise neural signals that control a user’s muscles and interpret them as commands to perform various tasks. These include typing or navigating through a user interface, without using traditional input tools.
Gestures supported using Meta’s sEMG wristband (Tap to expand)
Photo Credit: Meta
sEMG is non-invasive, and the company says that the wristband technology that it has developed can help users with “diverse physical abilities and characteristics”. Meta also claims that personalising the technology, using individual data, can help the machine learning models offer up to 16 percent better accuracy.
A video showcasing the technology on Meta’s blog demonstrates wrist-based one-dimensional cursor control while using the wearable. It shows a person writing the words “Hello World” on a flat surface using their index finger, without any stylus or input tool. It also shows a gesture that moves a cursor, as well as recognising certain discrete actions.
The evolution of Meta’s sEMG wristband (Tap to expand)
Photo Credit: Meta
In the past, Meta has open sourced its datasets with recordings for tasks like surface typing, and researchers can now access a new dataset that contains sEMG data performing three tasks, recorded by 300 research participants. The company also says the technology behind its wristband can offer better performance as it is used, by adapting to a person’s usage.
The company has also highlighted that sEMG input could help people with tremor to type or navigate a touchscreen interface. Meta has also partnered with educational institutions like Carnegie Mellon University, testing these technologies for persons with disabilities.