(Photo: Consumers need to either scan a QR code using their smartphones or pass through gates that use facial recognition technology to enter JD's X supermarket in Xi'an, Shaanxi province by Yuan Jingquan/For China Daily)
BEIJING -- Zhu Long remembers clearly how he got the first order for his company Yitu Technology, a Shanghai-based startup focusing on computer vision technology.
That was 2012 when Zhu, a postdoctoral fellow at the Massachusetts Institute of Technology, just returned to China. He chose the security industry as the starting point to build his own business.
But with no clients and without any background in industry, Zhu ran into trouble - until he managed to have a three-minute conversation with a deputy chief of a public security bureau in a Chinese city.
"He was too busy to listen to what I had prepared, and just said, 'Now our computer system can barely recognize 30 percent of false car registration plates. If you can boost the figure to over 70 percent, we will use your technology'," Zhu recalled.
Three months later, Yitu clinched the deal. Yitu offered a system that boasts recognition accuracy of 90 percent. Its success highlights how important technological superiority is to crack open a traditional sector dominated by established players.
"Though difficult to make the first step, the security industry is the native soil to nurture facial recognition technologies, for police departments often have a huge demand for real-time image-matching services, and they deal with tons of data that can help train our system," Zhu said.
That initial success helped Yitu to get a foothold in the multibillion dollar security industry.
To better penetrate the sector, Yitu forged a joint venture in 2016 with Synthesis, a firm that specializes in identity recognition solutions and boasts years of experience in the sector.
The joint venture will integrate Yitu's AI algorithms into Synthesis' hardware to offer better services.
Xiang Yang, an AI expert at the China Center for Information Industry Development, said application scenarios are of vital importance for AI companies, for differences such as light in diverse environments can affect the accuracy of facial recognition systems.
"So, the key is to have deep understanding of each application scenario, and come up with best solutions accordingly. Teaming up with experienced players is a way to quickly achieve that," Xiang said.
Yitu is also marching into other sectors such as finance, transportation and healthcare. In May, it raised 380 million yuan (60.35 million U.S. dollars) in its Series C round of funding, led by Yunfeng Capital, a private equity firm backed by Alibaba founder Jack Ma, and Sequoia Capital.
The company plans to set up a research and development hub in Singapore soon, its first such facility outside of China.
The hub is expected to bolster Singapore's applications of AI in the areas of security and finance. Singapore will likely act as a gateway for the startup to expand across the world.
"Singapore is a strategic location for us to begin our overseas ventures," Zhu said, adding Europe and Africa will be tapped next. (Source: China Daily)
A single purchase
