The startup beats Google with face recognition technology

The Russian startup NTechLab was founded only a year ago, but it has made headlines with its own facial recognition technology.

The company quickly became the leader in this emerging industry. In the MegaFace facial recognition contest held in Washington State last year, it defeated Google in one fell swoop.

The company has conducted more than 30 successful tests and received more than 300 orders. It is ready to bring its facial recognition technology to the world.

It plans to offer its cloud facial recognition system to customers such as businesses, governments and law enforcement.

This cloud service allows customers to upload a photo database and use it for facial recognition. Later this year, NTechLab will also launch a software development kit for third-party developers, and will also launch a factory security system powered by its facial recognition technology.

NTechLab differs from its competitors in the high accuracy of its facial recognition system and its power to search large photo databases. In the MegaFace contest, NTechLab achieved 73% accuracy in a database of 1 million photos; in a database of 10,000 photos, it achieved a jaw-dropping 95% accuracy.

“We took the lead in mastering the technology to efficiently process large photo databases,” said Artem Kukharenko, founder of NTechLab. “This advantage effectively helps us solve real-life problems, such as Search for criminals in real time, or identify frequent visitors from store surveillance cameras."

The startup has now 20 employees. It uses deep learning techniques and neural network structures. According to Cook Harenke, the hardest part of the process is the first facial recognition. This is the bottleneck in the whole process. The company's team is working to improve facial recognition algorithms to identify faces faster with fewer resources.

After scanning the face, the facial recognition system creates a feature vector containing 80 numbers and records the information of the face in detail. Everyone has a unique combination of numbers that can be used to distinguish them from others.

At this stage, the system also recognizes features that remain constant throughout the face; even when you get older, have a beard, or wear glasses or a baseball cap, these facial features will not change. In the final phase, the system compiles this information and uses them to search for matching information in the photo database. NTechLab's system can even retrieve ten times as many databases, and the retrieval speed will only be reduced a little.

NTechLab has tested its facial recognition technology in the crowd, at Australian amusement parks and at the Alpha Futures Music Festival in Russia this summer.

At the Alpha Futures Music Festival, the company invited people who went to the concert to provide their selfies to the NTechLab system. The NTechLab system scans and stores these photos and uses facial recognition technology to find someone in the crowd. When the search is successful and a matching person is found, the NTechLab system will send the matching photo directly to the person's mobile phone.

Earlier this year, another public showcase of its facial recognition technology, NTechLab also launched a free application FindFace. The app allows users to take a photo of someone on the street and submit it to the facial recognition system for identification. The system can successfully find out this person by searching the Russian social network Vkontakte and comparing the posts posted on this social network.

Packaging Staple

Zhejiang Best Nail Industrial Co., Ltd. , https://www.beststaple.com