The most common ways of collecting health data include devices like blood pressure measuring instruments and wearable devices and smartphones with APPs. Recently, a novel physiological-signals measurement technology is introduced by FaceHeart Inc., branched out from National Chiao Tung University. It requires only the images of human faces to measure heart rate, mood, and blood pressure. "so the discomfort of wearing devices can be avoided," said Meng-Liang Chung, the co-founder of FaceHeart Inc.

FaceHeart Inc. is founded by Prof. Bing-Fei Wu and his students in 2015 at the Institution of Electrical and Control Engineering at National Chiao Tung University, researching AI image recognition technology. They believe that the technology will become a trend and have great opportunities to combine with commercial applications, which is also the future vision of Prof. Wu.

Prof. Bing-Fei Wu at the Institution of Electrical and Control Engineering in National Chiao Tung University

In 2017, FaceHeart joined the Taiwan Startup Project run by the Ministry of Science and Technology. In the same year, FaceHeart attended the Longterm Cheers Startup Competition and won the second prize in fatigued driving. However,  Chung thought that an academic research team should be more ambitious in facing the market challenges, "in the market, only car manufacturers and field operators are interested in license plate recognition, and most drivers are hostile to the products related to fatigued driving detection." In Chung''s opinion, if they kept the focus on developing the license-plate recognition products, it would narrow the future of FaceHeart. Therefore, Chung and Prof. Wu reevaluated the technology''s potential and market needs and decided to devote to the medical field. In 2018, FaceHeart Inc. was officially founded and branched out from NCTU.

The foundation of FaceHeart Inc. is inspired by the fact that Taiwan is becoming an aging society. The healthcare industry may lack resources, and the tragedy of eldercare happens all the time. Even worse, physiological-signals measurement with a contact causes discomfort and inconvenience. For example, the compression of the blood pressure measurement devices may cause indisposed feelings or pains in elderlies'' arms. Therefore Prof. Wu and his team decided to establish FaceHeart Inc.

The big difference between campus and mall, marketing is more important than technology

Shifting from academics to markets and learning the difference between them are the challenges for FaceHeart. In academics, students need to focus on the research and the speculation of theories. Even though they have complete technologies, they still need to learn how to apply the technologies in the markets. Also, they need to consider finance and costs to make the company profitable and stable. Hence, Chung''s past experiences in startup play an important role.

Before establishing FaceHeart, Chung had started an undertaking with the main product of license plate recognition. As an inexperienced newcomer, he had been "scammed." Survived through the tough times, he learned a lot of precise experiences. Most importantly, he learned how to get inspired by the market needs for product design and come up with strategies from the angle of consumers, such as the pricing problem. "When it comes to the market opportunities, I am usually "against" my teacher. Then, brainstorming becomes practical business layouts. After all, running a company is something else," stated Chung.

Video recognition is not hard, but accuracy makes a difference!

The "Healthcare Management System with Physiological Signals Detection via Images" of FaceHeart can determine whether a person''s health condition is abnormal, based on analyzing the color changes of facial capillaries using images. When hearts pump blood, it generates P, Q, R, S waves, usually shown in the electrocardiogram, that indicate the heart''s beating. Among the waves, R waves are the easiest ones to observe. The system records the changes between each image in the forms of signal enhancement to measure the heartbeat, heart rhythm, mood, and anxiety level. It takes only six seconds for the measurement. As for the accuracy, Chung proudly claims that the errors are within 3bpm, which barely affects the measurement, compared to the data measured by hospital equipment.

Healthcare Management System with Physiological Signals Detection via Images

Knowing the technology''s market value, FaceHeart has passed the reviews of IRB (Institutional Review Board) of six hospitals and corporate with doctors for clinical validation of their technology. The system applies AI algorithms to construct a database to learn the different averages under various circumstances to improve data comparison accuracy. Moreover, their research team reduces the complexity of the AI matrices for edge computing in embedded systems so that the whole technology can be independent of cloud computing.

Wide range of applications: smart finance, healthcare, community safety, and more.

Another superiority of FaceHeart is that its software does not require customized cameras and can be run on any camera, including smartphones. The system can maintain high accuracy even with low-quality images. "VGA level is enough," said Chung. Therefore, they can apply their image recognition technology in elderlies'' care. For example, the system can be installed on the television and records the family''s physiological signals when they are watching the television. The system can also be applied to community safety for detecting suspicious people.

In 2019, FaceHeart received the A round of investment, leading by MediaTek. Inc, and started the new application of image recognition: finance technologies. Because of the increasing need for remote account opening, FaceHeart authorized Shanghai Commercial and Savings Bank to apply image recognition technology to identify the account openers and payees. Combining with the KYC (Know Your Customer) mechanism in the financial industry, the technology can be used for preventing financial crimes like fraud or money laundering. Moreover, sports measurement is also a potential market. With this measurement software installed in the cameras, users can learn their exercise status during running or spinning biking.

Application in smart finance

FaceHeart Inc. plans to include two functions into the system this year: respiration rate detection and blood pressure detection. Their business model consists of two types: software licensing and combining with camera products, targeting both enterprise and general consumers. The functions of the products will be enabled according to the using circumstances. For example, it only requires three software for home care: human face recognition, heart rate detection, and blood pressure detection, and the fee will be charged according to enabled software.

Dr. Meng-Liang Chung, the co-founder of FaceHeart Inc.