Could RADAR solve image privacy?
28-03-2021 | By Robin Mitchell
A recent announcement by Google shows that Google will use RADAR for gestures and determining to track sleep. What challenges do image-based systems face, what is Google developing, and could RADAR be the answer to privacy?
The Trouble with Cameras
The development of IoT and imagery technologies combined with AI has led to the creation of unimaginably powerful products. From automatic facial recognition to self-driving cars, visual technologies allow for a system to be aware of its surroundings and interact with them.
Therefore, it comes as no surprise that even the most basic applications now integrate some form of camera, internet connectivity, and processing capabilities. However, the sudden surge of IoT devices with camera capabilities has resulted in millions of potentially insecure devices accessible to criminals. Furthermore, these devices are often mounted in potentially private locations such as homes that can seriously affect users' privacy.
One method to get around such issues with pre-processing whereby the main processor of a design can only access imagery data that has already been preprocessed. For example, a camera designed to identify individuals could instead generate hash values for faces that are unique to each individual, and these hash values are passed to the main processor. Therefore, if a hacker gains entry into the system, they cannot see raw video, but instead, hash values in the area.
Google Announces use of RADAR in the Nest Hub
Recently, Google demonstrated the next generation of its control terminal, the Nest Hub. This device allows a user to control various aspects of their home from a single location including locking doors, turning on and off TVs, and controlling lights. However, the next generation of Nest Hub devices will also be repurposed as health-monitoring devices used in the bedroom.
The device will monitor users' sleep patterns using RADAR and do not require the user to wear any device. Furthermore, the device's use carries privacy concerns, so to address this the next generation of devices will not have a camera installed anywhere. Therefore, sleep patterns are detected with the use of RADAR only. Microphones in the system allow the device to monitor for snoring and coughing during sleep, but this does raise some concerns over privacy as some are known to talk in their sleep.
The RADAR system used by Google also allows for tracking motion, determining the number of people around the area, and detecting gestures that can perform specific commands. Additionally, the device is also designed to be a smart alarm clock and provides realistic features for providing a gentle cycle to wake up the user (i.e. gradual increase in brightness, change in colour, and slowly increased sound).
Could RADAR solve the privacy concern?
The RADAR system developed by Google demonstrates how cameras, albeit extremely useful, are not always necessary. If a camera can be removed from a design, it is most likely that user privacy is significantly improved.
Of course, RADAR cannot solve all problems. Still, as technology improves, designs may improve their resolution to provide detailed mapping without the need for cameras (such as that found in LiDAR), which may even be enough to recognise individuals.
Read More
