China’s computer vision company Megvii, a.k.a Face++ (旷视科技) has closed their Series C funding round at a staggering US$460 million. The investment was led by China’s state-owned venture capital Guo Feng Fund, joined by Alibaba’s Ant Financial and the Foxconn Group. Other investors include Russia-China Investment Fund (RCIF), Sunshine Insurance Group, and SK Group. Tengda Capital is the exclusive financial advisor for the deal.
Face++, established in 2011, received its A round financing from Legend Star Capital. It’s been a rich year for Chinese AI startups — computer vision company YITU closed their C round investment at US$380 million this May, while SenseTime’s B round financing totaled US$410 million.
Qi Yin, the company’s 28-year-old Co-Founder and CEO, graduated from both Tsinghua and Columbia University’s computer science programs. He tells Synced that Face++’s strategy is to focus on facial and image recognition and forging AI brains. For image recognition, Face++’s AI extracts information like facial features, vehicles, and semantic information, which are applicable in IoT, transportation, and fintech for building future smart cities.
Face++ will be building on its forte in finance, security, other niche areas like real-estate. The company is also a dark-horse in the mobile application race, providing technology for Xiaomi Note 3, VIVO V7 Plus’ face unlock feature and AliPay’s facial recognition payment service. Face++ has been used 20 billion times on 500 million devices. It also helps Chinese police in 25 provinces, so far participating in over 1,000 crime cases and the arrest of some 2,000 fugitives. The platform also has an impressive 210 million verified-name passenger database for subway, airports, and railway travel.
Face++ holds 440 patents in China and abroad. As one of China’s earliest AI startup, the company’s deep learning engine Brain++ is a driving force in engineering, product development, and industrialization. Face++’s open-source developer platform has attracted over 30 million users.
Qi attributes Face++’s success to what the company calls its “4 in 1 formula,” which combines algorithms, software, hardware and data in a single solution package; and the “0 or 1 formula,” used for managing data. “If an AI company does not have its own closed-loop datasets, they can only get short-term profit. Obtaining core data is crucial for AI infiltration into specific industries.”
On July 4 of this year Face++ introduced ShuffleNet, a computation-efficient CNN architecture for mobile computing with the power of 10-150 MFLOPs. The model reduces computational complexity but maintains high recognition accuracy. By introducing the innovative pointwise group convolution and channel shuffle technique, ShuffleNet reached an accuracy level 6.7% higher than Google’s MobileNet at 40 MFLOPs.
“AI will provide big returns, but we need to understand that the technology develops in many iterations,” says Qi, whose eventual goal is to make Face++ a general solution AI platform that can penetrate and monopolize core industries and change the business landscape.
Journalist: Meghan Han | Editor: Michael Sarazen