System and method for detecting, recording and communicating events in the care and treatment of cognitively impaired persons

A system for detecting and analyzing fall events at a particular care facility comprising:

• a server and a first data storage system remote from the particular managed care facility constructed to aggregate data from a plurality of managed care communities;
• to train at least one convolutional neural network from the aggregate;
• and to distribute said convolutional neural network to the particular managed care facility;
• a 2-dimensional camera located at the particular care facility, the camera assigned to a particular room in said care facility, the camera having a field of view, the camera activating 2-dimensional video recording upon detecting motion in its field of view;
• the 2-dimensional camera producing a time sequence of video frames;
• a fall event detector constructed to detect a fall event from a recorded sequence of video frames originating at the 2-dimensional camera, the fall event detector creating a fall alert upon detection of a fall event, including fall event information;
• the fall event detector including said convolutional neural network trained from the aggregate data from the plurality of managed care communities;
• the fall event detector configured to produce a final classification and detection of the fall event based on both pattern recognition and on the time sequence of video frames;
• wherein, a detection score is used to determine the detection of a fall event, the detection score increasing with larger frame-to-frame pattern changes;
• the fall event detector transferring said fall alert, said fall event information and the recorded sequence of video frames relating to the fall event over a communications network to the remote server, the remote server constructed to receive notification of event validity from an event analysis service upon receiving said fall event information and the recorded sequence of video frames;
• wherein, the recorded sequence of video frames are presented a the analysis service as a video sequence capable of being verified by a human observer;
• the remote server constructed to store said fall alert, said fall event information and said recorded sequence of video frames on a second data storage system.

A system for detecting and analyzing events at a care facility caring for patients with cognitive impairment comprising:

• at least one single 2-dimensional camera located at the care facility, the at least one single camera assigned to a particular location in said care facility, the camera having a field of view, the camera activating 2-dimensional video recording upon detecting motion in its field of view;
• an event-detecting processor configured to detect events of interest from a recorded sequence of video frames originating at the single camera, the event-detecting processor configured to detect said events, and the event-detecting processor adapted to create an event alert upon detection of an event, including event information;
• the event-detecting processor including a convolutional neural network trained from data collected from a large plurality of managed care communities;
• the event-detecting processor configured to produce a final classification and detection of an event based on both pattern recognition and on the time sequence of video frames;
• the event-detecting processor transferring said alert, said event information and the recorded sequence of video frames relating to the event over a communications network to a remote server, the remote server constructed to receive notification of event validity upon receiving said event information and the recorded sequence of video frames;
• the remote server constructed to store said alert, said event information and said recorded sequence of video frames on a storage device;
• a post-processing process wherein the sequence of video frames is analyzed to 1) understand how events of interest affect staff and patient safety and 2) provide training to prevent said events of interest.