Surveillance Neural Networks

A new consumer surveillance system utilizes neural networks to better identify who or what is approaching your home. So instead of poring over hours of video that was recorded over the course of the entire day, users are presented only with the events that contain potentially significant information.

Unlike currently available surveillance camera systems that create alerts based solely on motion, sound or face recognition, Camio uses artificial intelligence and machine learning to search out specific objects with greater precision. Information from multiple cameras can be compared concurrently to determine whether an event is considered significant. As users review the selected clips, they help the system learn by indicating whether or not a particular clip was important. Eventually the amount of footage to be reviewed can be honed down to about a minute a day, which can be stored on a remote server.

Users can also set up alerts based on an “If This, Then That” (IFTTT) response-triggering service. For example, the system can be pre-programmed to alert your smartphone whenever a package is delivered. Camio allows a variety of smartphones and tablets to be used as surveillance cameras, and works with some individual cameras as well. Users may stream video for free, but the company charges $9 per month per camera for recording, playback and storage.