Sang Won Lee, founder and CEO of Qishou Company
Tencent's "Front Line" Puxiang
Qeexo Qishou has licensed FingerSense to Huawei and EarSense to OPPO. Now it is displaying its embedded machine learning products from CES to MWC2019 to find more application scenarios in the field of IoT.
Qeexo's machine learning platform has entered more than 180 million smartphones and tablets worldwide through products such as FingerSense and EarSense.
In Soho's office in Wangjing, Beijing, the author of Tencent's "Front Line" met Sang Won Lee, founder and CEO of Qishou Company. He is a young Korean, in English, said that since the company was founded in 2012, the number of team members has not exceeded 60, mostly engineers.
Lean teams and a large number of markets make this AI company relatively prosperous. After the establishment of offices in Beijing and Shanghai, Qishou set up offices in Shenzhen in 2018.
Qeexo embedded machine learning is a lightweight universal platform, which can perform machine learning reasoning locally on embedded edge devices in real time without relying on the cloud.
In industrial environments, Qeexo-driven sensors can be set up in factories to monitor and analyze processes, equipment, and target products, enabling machines to run in a more ideal state for longer periods of time. In the automotive industry, if the sensor has Qeexo embedded machine learning, it can give immediate feedback to the road and car conditions, or predict the maintenance of the car. In the field of smart home and Internet of Things, edge devices can expand more useful and convenient functions with lower additional prices.
On the basis of smartphone business, Qishou wants to enter a broader business area. Qeexo embedded machine learning can make products and processes more intelligent in many areas. It can achieve millisecond delay, which makes the application feel like instant feedback.
And this platform has very low power consumption, memory and processing requirements. Qeexo embedded machine learning is highly optimized to allow reasoning on edge devices, and has a wider range of applications. It works for all types of sensor data.
How to get more customers in the field of IoT is a new challenge for Sang Won Lee and his team.
Sang Won Lee, an MBA student at the University of Berkeley, worked for Samsung, looking for innovative smartphone interaction experiences. Once, when he read Chris Harrison's academic paper on AI interaction, he contacted Chris on his own initiative and finally convinced him to start a business together.
Since then, in order to find users for the product, smartphone application scenarios have been found among many options. Since then, Sang has spent a long time on the mainland of China. He said it didn't take much effort to win over Huawei, a Chinese customer.
In 2014, Huawei introduced the simple operation of finger joint screenshots to users as a highlight. This operation is much easier than pressing both the power supply and the Home button on the phone.
The operation of FingerSense is very simple. If you bend your index finger or middle finger, tap the screen two times with the second finger joint, the mobile phone can complete the screenshot operation. In addition, finger joint gestures can help to achieve rolling screenshots, regional screenshots, video recordings, split screens, and can also quickly open applications. For example, just tap the screen with your knuckles and draw the letter m, and the music application will open.
As an artificial intelligence technology, FingerSense has strong adaptability. It does not need to add sensors and other hardware in smartphones, but relies entirely on software to achieve it. It requires very little memory and energy consumption, so that mobile phone manufacturers do not need to worry about other functions being affected.
After FingerSense, Qeexo has introduced two new technologies: TouchTools, a gesture virtual tool for large screen size, and EarSense, an ear-sensitive breath-screen technology that replaces proximity sensors.
Among them, in June 2018, EarSense technology was applied to OPPO flagship Find X full-screen mobile phone. The technology recognizes the proximity or touch of a mobile device to the face and ear, and turns off the screen when the user answers the phone. Sang believes that in the current trend of Liu Haiping, this technology breaks the limitations of traditional short-range sensors and saves the upper border space for mobile phones in design, thus making the real full screen possible.
Previously, there have been some solutions to replace traditional short-range sensors, but these solutions need to replace the old hardware with new hardware, which has high cost and general performance. EarSense, by contrast, is a pure software solution. Mobile phones no longer need to reserve a location for sensors, so that mobile phone engineers have more space to play.
Sang was using Find X when he met the author of Tencent's Front Line. According to OPPO, Find X eventually sold nearly 2 million units.
Of course, IoT field experts such as Lin, how far can Qi hand go in the future remains to be answered.