Contactless Medical Event Detection via Deep Learning and Biometric Sensors


No devices or technology to wear or use.

How comfortable are your parents or grandparents using technology?


No devices or technology to wear or use.

Come on, Seriously?

Look at all the technology and devices that the current market leaders have your parents using. .. All of this technology may look good on a business plan or in a brochure but the reality is that all of this stuff will rarely if ever get used by the elderly.  In fact, the British Medical Journal proved what family members already know: many older adults do not wear the pendant bought by their adult children, typically not wanting to appear old. And if worn, most do not press the button, not wanting to bother other people.  Below is a partial list of notable limitations of current Medical Alert / Personal Emergency Response Systems:

     1: The user fails to activate the alarm

     2: The user fails to provide a status after activating the alarm

     3: The user is embarrassed to wear or use the pendant

     4: False Alarms done on purpose because of loneliness or depression


Learn More about The Limitations

Assisted Living Suite Design

Deep Learning paired with unobtrusive environmentally embedded biometric sensor technology will be installed in living quarters and public spaces to automatically monitor and detect rapid (acute) changes in health or functional status. If an acute medical event has occurred then an automatic alert will be sent to the users’ care provider for urgent response and treatment. The goal is to reduce hospital admissions by enabling an immediate and educated response to acute medical events which in turn will minimize the time spent unattended on the ground (long lie). The longer the user spends unattended on the floor the greater the probability that the severity of the injury increases. Patients have No Technology or Devices to Use or Wear.

Privacy Concerns

Privacy is one of the most important topics for MEDANN to address and communicate since it’s usually the first area of concern when people learn that MEDANN will be using cameras. Our first response is to immediately clarify that we will not be using a video camera in the traditional sense. Our cameras are closer to sensors than video capturers.  MEDANN camera/sensors monitor using machine vision and deep learning. This means that our technology will summarize the situation as opposed to specifically watching activities of individuals: no humans are involved in our automated monitoring process thus now giving customers the option of using a privacy-preserving monitoring technology.

MEDANN will use skeleton (stick) mapping and background subtraction. Subtraction extracts regions of interest by comparing the current frame information with the background model.  All that will be seen by administrators and staff is the machine vision images as seen to the left.