Sighthound achieves 99.79% accuracy against LFW benchmark, beating Google, Baidu and Facebook, makes system available for free to developers on the Sighthound Cloud service.
Palo Alto, California – December 16, 2015 - Sighthound, Inc. today announced that Sighthound’s new facial recognition system ranks first against the Labeled Faces in the Wild (LFW) benchmark database hosted by the University of Massachusetts, with an accuracy of 99.79%. Previous highest accuracy scores include Google at 99.63% and Baidu at 99.77%. Sighthound also announced that it is making the system available to developers for free on the Sighthound Cloud service. Sighthound Cloud also offers face detection, person detection, gender recognition and facial landmark identification.
Founded in 2012, Sighthound develops innovative proprietary computer vision systems and is a leader in deep learning for real world implementations. Sighthound’s facial recognition system stands out by virtue of using less than 2% of the amount of training data used by Google, and by using only one crop per image. Some others on the LFW results list use up to 25 crops of the same image, each crop showing a small part of the face, which slows down or prohibits real world applications. Further, Sighthound’s algorithms are robust to both verification tasks such as LFW and identification tasks such as tests against the PubFig200 database. The result is that Sighthound’s software is highly accurate, robust to a variety of real world use cases and runs in real time.
“We have spent 3 years quietly working on the commercial viability of deep learning and computer vision,” said Stephen Neish, Sighthound’s CEO. “These results are testament to the caliber of PhDs we hire and our focus on real-time, real world uses of our research. In this case we were forced to build more intelligent deeply learned networks because we don’t have access to the amount of data that Google or Baidu have.”
Sighthound’s new facial recognition system adds to the set of developer APIs available on Sighthound Cloud, and joins an expanding range of products that use Sighthound’s state-of-the-art computer vision algorithms. Sighthound also markets Sighthound Sentry, the embedded computer vision engine used by OEMs to add intelligence to their cameras for the home automation and security markets, and Sighthound Video, advanced computer vision security software for homes and businesses.
Sighthound makes sense of movement through our world. By building state-of-the-art computer vision research into easy-to-use software, Sighthound makes it easy for computers to see.
Sighthound: Rainer Flor, Sighthound, firstname.lastname@example.org