A startup that licenses Purdue University technology has created a camera and hardware package that can be taught to recognize objects or perform tasks that could be used in security, automotive, industrial and defense applications.

Current [camera] systems have problems distinguishing between things such as humans and pets because they detect only movement, instead of a particular object or person, says Eugenio Culurciello, associate professor in Purdue's Weldon School of Biomedical Engineering and co-founder of TeraDeep. These systems have a large number of “false alarms: and have to record huge amounts of video to determine if there is a threat, which wastes time and resources," he says.

The TeraDeep system records segments only when a person or object that the client wants to identify is present. It also can allow for smart locks to give people access to different locations based on facial recognition.

The camera can allow for smart locks, giving people access to locations based on facial recognition. Image credit: TeraDeep.The technology can be trained to perform by watching similar videos or being shown examples. Users can apply rules to enable it to detect an action or object and recognize individual faces.

TeraDeep uses its own hardware system so footage can be processed faster with more privacy than conventional methods, the researchers say.

"Video that is captured on current camera systems usually is streamed in the cloud, and people hosting the server look at the video and process the parts that are interesting, so users don't have to look at hours of footage," Culurciello says. TeraDeep uses an accelerator-embedded hardware system, which means footage can be processed locally, privately and more quickly.

Culurciello says the technology gives machines the ability to see the environment around them, which could make it useful to many industries that need to process large numbers of images or videos.

For example, the system potentially could be used in automobiles to perceive the outside environment and take action based on identifying certain elements, such as pedestrians, to navigate safely and securely.

Another potential application involves changing how a device performs based on who is operating it, for example, switching television settings based on whether a child or an adult is watching, or automatically adjusting the temperature based on who is in a room.

Culurciello says TeraDeep has already received interest from the defense and automotive industries.