New 3D sensors track body posture with wristband, integrate processing
Researchers from Cornell University have developed a wristband capable of tracking an entire body’s posture in 3D.
The device, called BodiTrak, relies on a deep neural network that allows a wrist-mounted camera to produce a full-body graphic of the wearer’s actions in real time.
Spotted by Technology, the wearable solution could one day be integrated into commercial watches and phones. It was described in a research article published by the Association for Computing Machinery.
“Since smartwatches already have a camera, technology like BodiTrak could understand the user’s pose and give real-time feedback,” says the paper’s senior author, Cheng Zhang. “That’s handy, affordable and does not limit the user’s moving area.”
The deep learning model can estimate the 3D positions of 14 body joints. It was tested on nine participants performing 12 daily activities including walking.
The tool has an average error of 6.9cm (2.7 inches) using one 11.5mm by 9.5mm red-green-blue camera on a wrist pointing toward the body, reads the paper.
The model can read crude images or silhouettes of the user’s body in motion and combines them with data captured by a sensor.
“Our research shows that we don’t need our body frames to be fully within camera view for body sensing,” says Hyunchul Lim, a doctoral student and the paper’s primary author. “If we are able to capture just a part of our bodies, that is a lot of information to infer to reconstruct the full body.”
According to the study, watches do not have cameras and sufficient battery life to integrate full-body sensing.
Blickfeld launches lidar sensor
Munich-based lidar system maker Blickfeld announced the availability of a new sensor that captures and processes 3D data on a single device.
Called Qb2, the sensor was unveiled at CES 2022 and is now available for delivery.
The integrated compute module enables the company’s Percept software stack to run on the device, providing data for volume monitoring, crowd analytics and other sensor infrastructure.
By removing the need for additional computers and bringing faster insights, Qb2 also improves response times and provides improved bandwidth availability via edge computing.
From a technical standpoint, Qb2 also offers improved functions and performance such as enhanced adaptable field of view, detection range, scan pattern and resolution.
One of the applications for lidar technology is high-performance biometrics.
The tool’s release comes almost a year after Blickfeld secured funding in a $31 million series A valued.
According to Crunchbase, the company raised a further €15 million (roughly $15.27 million) in debt financing last July.