Original Title: Face Era, Face Recognition for Entrepreneurs to Break Through the Road
On September 26th, a student entered a simulated exam room in a Beijing middle school after confirming their identity through a face recognition system. The photo was taken by Beijing News reporter Wu Jiangshe.
Apple has always been a pioneer in driving technological change across industries. At an Apple event in early September, “face recognition†once again took center stage. The new full-screen iPhone uses Face ID technology, sparking excitement among Chinese users and making many blush with admiration.
From "face scanning to collect housing funds" to "face scanning for attendance" and even "face scanning in public restrooms in Ritan Park," where toilet paper is dispensed only after a facial scan—more and more daily activities are now associated with face recognition. This technology is becoming increasingly embedded in our everyday lives.
When combined with government services, face recognition could soon make it unnecessary to carry around multiple documents or fill out lengthy forms. A simple swipe of your face might be all you need to prove your identity in the future.
The growing interest in face recognition is also reflected in the funding of related companies. Face++, a leading cloud-based face recognition platform, and Shangtang Technology, which specializes in AI vision engines, have both secured significant investments—$100 million in Series C funding and 410 million yuan in a B-round respectively. These companies are among the "unicorns" in the AI space.
It seems that the tipping point for commercial applications of face recognition is finally here. While the iPhone X may not have fully realized its potential, it did bring face recognition into the mainstream.
Three Face Recognition Algorithms
What exactly is face recognition? In short, it's a biometric technique used to identify individuals based on facial features. It typically involves three steps: image acquisition, feature detection, and identity verification. Essentially, the system extracts key facial features like bone structure, eyebrow height, and compares them to determine identity.
Although Apple’s Face ID has generated excitement among computer vision entrepreneurs, it is not the same as traditional face recognition. Apple’s system uses infrared technology instead of regular cameras, enabling 3D recognition and significantly improving security.
In practical applications, face recognition can be divided into three main types: 1:1, 1:N, and N:N.
1:1 face recognition is the most basic level, used to verify "you are you." Users upload a photo in advance, and during each verification, the system compares it with the stored image to confirm identity. This method is highly accurate and requires less computational power.
For example, when passing through a station’s security checkpoint, the staff checks your ID against your face to confirm you're the rightful owner. This is a classic 1:1 scenario. Applications include phone unlocking, face payments, online ticketing, hospital registration, government projects, and more.
1:N face recognition, on the other hand, is used for identifying "who you are." It compares a single image against a large database of faces and ranks the results by similarity. This type is often used in security contexts, such as tracking criminals or locating missing children.
One startup, Yun Tian Li Fei, collaborated with the Longgang District Police in Shenzhen in 2015 to implement a "Deep Vision" system at subway stations and railway hubs. Within months, the system helped solve two murder cases.
1:N recognition is dynamic and non-cooperative. The system captures video footage rather than still images, and the subject doesn’t need to be aware or cooperate. While convenient, this also means your location and movements can be tracked without your knowledge.
Compared to 1:1 recognition, 1:N is more challenging due to factors like lighting, angles, and environmental conditions that can affect accuracy.
N:N face recognition is essentially multiple 1:N identifications happening simultaneously, used to verify identities in complex scenarios.
Face Recognition Technology Is Not Yet Perfect
While China, Europe, and the U.S. are on par in basic research, China leads in commercial applications. Experts from Beijing University of Aeronautics and Astronautics note that big tech companies in the West focus more on long-term goals, while Chinese firms prioritize quick AI implementation through face recognition.
Many AI entrepreneurs in China are competing fiercely. From internet giants like Alibaba and Tencent to unicorns like Face++ and Shangtang, face recognition is no longer exclusive to big companies—it's becoming a common tool in the market.
However, the real challenge lies in real-world application. As investor Zhang Quanling notes, the threshold for face recognition is low, but implementing it successfully in real-life scenarios is far from simple.
In a recent TV show, a deep-learning robot demonstrated current capabilities: it can extract thousands of facial features and match them in seconds. But challenges remain—poor lighting, angles, masks, or even digital alterations can all lead to errors.
Last year, during a live demonstration, a presenter used a 3D model of a viewer’s face to bypass a system, highlighting vulnerabilities in current face authentication methods.
Experts like Qiu Xuejun from 360 AI Research Institute warn that face recognition systems still lack robust liveness detection. Many use simple algorithms that are easy to fool.
The Scene Is the Key to Success
Many startups fall into the trap of believing that technical superiority alone is enough. According to PreAngel’s Jiang Jiang, most early-stage face recognition companies still focus on their technology without considering real-world applications.
As Zini Fund partner Zhang Hongjiang puts it, companies without data, strong applications, or clear scenes won’t last long. While data is important, it’s not the only factor. Small improvements in accuracy don’t always translate to better user experiences.
Despite this, technology remains crucial. Only when a company reaches unicorn status does technology become a major barrier for new entrants.
Investors like Jiang Wei emphasize that they’re looking for startups that can solve real problems, not just those with advanced algorithms. Commercialization capability is key for emerging companies.
Giants have clear advantages in resources, data, and capital. But startups can still thrive by deeply understanding and serving specific industries. As one anonymous investor noted, big players’ investments and innovations create new opportunities for niche markets.
Whether it's in entertainment or finance, the success of face recognition depends heavily on the context. Investors are more optimistic about scenarios with lower accuracy demands, like retail and leisure, and cautious in high-stakes areas like financial transactions.
Wang Jun of Yuntian Lifei agrees: "The most valuable technology isn't necessarily the most advanced, but the one that can be widely adopted." He sees great potential in areas where safety requirements aren’t as strict.
Beijing News reporter Cai Haoshuang
For the 132kV Oil Immersed Power Transformer, we can produce capacity upto 180MVA. We use the best quality of raw material and advance design software to provide low noise, low losses, low partial discharge and high short-circuit impedance for power transformer.
Our power transformer are widely used in national grid, city grid, rural grid, power plant, industrial and mining enterprise, and petrochemical industry.
132Kv Transformer,132Kv Power Transformer,132Kv Power Transformer With Oltc,132Kv Transformer With Oltc
Hangzhou Qiantang River Electric Group Co., Ltd.(QRE) , https://www.qretransformer.com