Amazon has patented facial recognition technology that could create a neighborhood network for suspicious persons. Facial recognition could be used to compare subjects to a criminal or “most wanted” database, and could allow users to access suspicious persons’ identifying information.
Amazon’s patent, filed in May of 2018, will allow users to “flag” suspicious persons like someone breaking a neighbor’s window or a hostile co-worker. Aside from what their cameras have captured, users can upload images of the person from other sources. From there, users can receive alerts about the suspicious persons, even if they’re not in any criminal databases. On the other hand, users can flag people as authorized persons who can also be added to the neighborhood database.
Aside from identifying people as suspicious or authorized, the patent allows cameras to recognize suspicious behavior. Algorithms can account for factors like loitering times and the software can recognize suspicious activities like concealing objects, running, or crouching. Subjects’ behavior can be compared to databases for further identification.
Even if the user believes the composite facial image is not recognizable, the software could compare the image to a database of suspicious persons to determine whether or not the suspect is a security threat. By comparing the image to the database,
“…the user of the client device may receive another layer of information about the person that may enable them to make a more educated decision as to whether or not to pursue law enforcement help, report the incident to their landlord, share the information with their neighborhood, etc,”
according to the patent. Once the camera detects a person, users can receive mobile notifications with audio/video data and from there, can decide whether or not to take further action. On top of that, a suspicious person can trigger an audible alarm or announcement from the camera, making further criminal activity difficult. The patent uses artificial intelligence to build upon neighborhood watch apps where users can manually upload photos or videos of what they deem to be “suspicious activity.”
Some are worried about how a neighborhood database of suspicious persons could unfairly target people who have been wrongfully arrested or imprisoned. In the same way, the software would cost a negative light on anyone with a criminal background, regardless of how long ago it was, what their actual crime was, or what they have done with their lives since. The database would be a digital cloud over anyone with a criminal past, and would make all of their neighbors aware of their wrongdoings.
To learn more about the best security cameras with facial recognition on the market today, read the best cameras for artificial intelligence.