Time-of-flight cameras make patient monitoring secure and private

Patient monitoring is essential to healthcare, but as the public grows increasingly sensitive to privacy, the use of video for this purpose is being questioned. Time-of-flight (ToF) cameras such as Omron’s B5L 3D ToF sensor module are a promising alternative.
These sensors inherently ensure patient anonymity as the sensor gathers raw distance data rather than image data. The body detection algorithm also allows for machine learning that can alert users when an incident is detected, thus removing the need for staff monitoring of feeds. The raw distance data gathered by the sensor ensures user security, as no compromising information can be collected.
Patient privacy is an urgent concern
For many years, video has been used to securely monitor patient rooms and other areas of healthcare facilities. Cameras can capture the movements of patients in great detail, but, ironically, it’s the technology’s high-resolution capability that is its Achilles heel. If recordings are stored, they can also be stolen either by someone in the facility or over its network, making it inherently insecure.
In the U.S, because the Health Insurance Portability and Accountability Act (HIPAA) and the Electronic Communications Privacy Act limit how personal data must be treated, video monitoring, storage and retrieval may not meet the requirements of most healthcare facilities. Another limiting factor is that effective video monitoring requires someone to constantly monitor multiple patient rooms on screen. For nurses, this adds to already staggering workloads and limits productivity.
What’s increasingly required is a technological solution that ensures patient anonymity, operates autonomously and provides a high level of accuracy under any circumstances. Thanks to advances in their capabilities, TOF cameras can now meet all these requirements at relatively low cost.
What is time-of-flight?
TOF is a rather amorphous term that can be applied to any sensor technology, from low-frequency ultrasound to radar and lidar, which works by transmitting a signal from a sensor that bounces off a target and returns to its origin embedded with useful information, primarily the distance between the source and target (see Figure 1). A TOF camera operates at optical wavelengths such as 940 nm that are not visible to the naked eye and uses an LED sensor as the light source. As the signal’s bandwidth is extremely wide, it can capture the details of a wide area in three dimensions in a span of a single light pulse.
The TOF concept itself was first defined mathematically more than 100 years ago, but over the years the technology to implement it at optical wavelengths has dramatically improved, especially since LEDs became available that enable TOF cameras to be made inexpensively. The technology now serves many applications from 3D mapping to industrial automation, obstacle detection, agriculture, robotics, measuring volumes and dozens more. Some smartphones also use TOF to provide location information and distance measurements.
On first appearance, TOF images would seem to be of little use because they deliver none of the detail of a camera. For patient monitoring, this is precisely what the doctor ordered. In short, the patient remains anonymous because the image does not provide facial recognition or other identifiable characteristics, so from a privacy perspective, there is nothing even worth collecting.
TOF camera: Low resolution, but in three dimensions
An image taken by a standard camera (first) captures images with great detail while the images from the TOF camera (second) are of low resolution but in three dimensions.
There is no possibility of identifying a patient's face or other features. (Source: Omron)
When combined with machine learning algorithms, TOF-camera-based systems can be trained to recognize normal activities for specific a patient over time, providing the ability to detect movements that the patient should not be making and automatically report these issues instantaneously. If this anonymous information is stored and updated, a record of activity can be created that shows how much time a patient is sitting up, moving around or lying (or falling) down.
They can even depict an entire person even when the entire body is not visible if, for example, the patient moves behind a bed. Omron’s B5L 3-D TOF sensor module achieves this using skeletal movement techniques that identify body parts from the return signal and assign markers of different colors to joint locations such as the knee, elbow and shoulder. As a result, when the person moves behind an object and the lower part of the body is obscured, the system uses its learned information to show the entire body. The system stores this information and updates it through machine learning algorithms to improve its accuracy over time.
Like any technology, TOF has its limitations. Foremost is its sensitivity to extraneous light that was a significant drawback until recently. However, the B5L 3D compensates for this with advanced optical techniques and software, providing immunity up to 100,000 lx, the equivalent of direct sunlight. The software also compensates for interference from other light sources, enabling it to operate effectively in any lighting environment.
The camera’s resolution is about 3 feet, and the capture area is 70 x 90 degrees over a distance from 1.5 to 20 feet, and the system software provides a variety of operating modes that can be involved to meet the needs of specific environments. In addition, it allows up to 17 sensor modules to be used in the same room by slighting changing the operating frequency in each unit.
In short, it’s not surprising that the acceptance of TOF for healthcare applications is growing substantially just as privacy and security issues become a paramount concern. TOF cameras such as Omron’s B5L eliminate these issues and, when combined with machine learning techniques, provide capabilities such alert hospital staff only when necessary.
Learn about Omron’s B5L technology and TOF sensor camera module.

