Facial recognition in the dark
- By Matt Leonard
- Apr 20, 2018
Facial recognition generally requires well-lit photos to make a match, but warfighters often need to identify individuals from images taken by thermal cameras in low-light or nighttime conditions. With the help of artificial intelligence, Scientists with the Army Research Lab have developed a way to enhance images from night-vision cameras so that they can be matched to conventional images in existing biometric face databases or watch lists.
The approach creates a visible facial image that has been synthesized from multiple regions of a thermal image. By combining features from the across the entire face with those of specific regions -- like the eyes, nose and mouth – the researchers were able to synthesize a refined visible image that could be run against a face database
Making a match between the visible and thermal image types required training a neural network to learn when the features from the two types of images were highly correlated, making it then possible for a matching system trained with visible imagery to be applied to thermal imagery.
This approach will allow the military to expand its facial recognition capabilities without developing custom software.
Read the full paper here.
Matt Leonard is a reporter/producer at GCN.
Before joining GCN, Leonard worked as a local reporter for The Smithfield Times in southeastern Virginia. In his time there he wrote about town council meetings, local crime and what to do if a beaver dam floods your back yard. Over the last few years, he has spent time at The Commonwealth Times, The Denver Post and WTVR-CBS 6. He is a graduate of Virginia Commonwealth University, where he received the faculty award for print and online journalism.
Leonard can be contacted at firstname.lastname@example.org or follow him on Twitter @Matt_Lnrd.
Click here for previous articles by Leonard.