Free facial recognition up to snuff with specialist options
Research has shown that freely available facial recognition search engines used by popular social networks like Facebook are as reliable and accurate as some specialist systems sold to government agencies, Police Oracle reports.
To conduct the study, a large mass of facial images were retrieved from Facebook, Face.com (acquired by Facebook) and Picasa — three major online services that rely heavily on recognition software – and measured how well each service compared in terms of detection rate, grouping and tagging.
“This wasn’t just a case of us throwing data at Facebook and seeing what comes back,” Thieme said. “We actually looked at every match that was returned and made sure it was correct. It was extremely labour intensive, but it was the only way for us to accurately measure how well these systems performed.”
The results of the tests shows that the social networks were extremely accurate in terms of facial recognition, particularly since images were not necessarily taken frontally, and in many cases, had other objects in frame that could have spoofed the software.
According to the article, Thieme says that to his knowledge, no one has measured this before and it contradicts the prevailing view that free or cheap tools are far inferior to sophisticated biometric tools produced by specialist companies.
Thieme suggests that for agencies that need to conduct a search using a small list of images, these free online services would be accurate enough for their requirements, and could even serve a better purpose.
“The concept that the only way to do reliable face matching is by using an enterprise technology with enterprise licenses is not correct,” Thieme said.
Thieme’s findings, along with many recent implementations of facial recognition technologies, make it clear this is a technology that’s increasingly accessible. Let us know: Are you comfortable publishing images of yourself online, knowing that powerful facial recognition technology exists widely in social networks?