No more hooded: facial recognition system detects covered faces

in technology •  7 years ago 

With facial recognition systems, it is easy to find the identity of a person. Thanks to new smart software it will be possible to identify someone even when their face is covered.

A group of technology experts presented a new intelligent computational algorithm capable of identifying the hooded by an advanced system. This makes it an excellent tool to fight against crime and vandalism.

Smart Face Recognition for Covered Faces

The cameras located in public places and shops capture images of what happens and, reviewing the videos, facial recognition applications can give the identity of those who appear in them.

The big problem was to identify those who wear their faces covered with hoods or handkerchiefs. We say "era," as American and Indian experts created an intelligent facial recognition system that works even with half-sided faces.

This new system performs Disguised Face Identification (DFI) or Disguised Face Identification. For this, the software is fixed in 14 key points of the face, 10 of them in the eyes and nose, which are usually discovered in the hooded.

Detecting these key points, the algorithm compares them with photographs already stored of the people and thus completes the faces to find the identity of who tries to hide for burglaries and disorders.

This program uses a type of artificial intelligence known as deep convolutional neural networks, which function similarly to neurons in the visual cortex of the brain. Thus, they can recognize images by parts and then process them to give a total according to the information collected.

The system created by the international group of experts compares the specific features collected with a series of images until one matches the one with the most possible points.

While the progress in DFI is encouraging, the results are still far from perfect, as in the case of photographs of people with hats and scarves on the face, with the correct identity in just over half of the cases, which drops to 43% if glasses are also included.

Also, the success or failure of this intelligent identification software depends on the light conditions in which the image was captured and, of course, on having a face database broad enough to be useful.

Smart facial recognition systems for covered faces will undoubtedly be of great use in the near future, as they allow the identification of those who make disorders that have hitherto remained unpunished.

follow.gif

Authors get paid when people like you upvote their post.
If you enjoyed what you read here, create your account today and start earning FREE STEEM!