A number of companies around the world are working hard to develop AI technologies and are opening an “era of artificial intelligence.” In Korea, many companies are also focusing on developing AI. CEO Jung Bae Kim and CTO Hwang Young Kyoo founded Alchera Inc. This is because there were no companies that had their own competitive edge in South Korea’s artificial intelligence field. The two are the main players in the Korean market, leading the commercialization of the technology while developing AI image recognition for more than a decade at the Samsung Advanced Institute of Technology. They started Alchera with full support from Naver, and commercialized four products in three years. Insight met the two at Alchera’s office in the Pangyo Techno Valley on May 22.
Alchera’s office was busy on the day of the interview. Coincidentally, it was the day when Alchera’s office, which was only using the eighth floor, expanded to also use the sixth floor. “I think cooperation is important. So we are trying to create an atmosphere that is horizontal and accountable,” Kim said.
Q. Tell me about Alchera.
Jung Bae Kim: Alchera was established in June 2016 as a company specializing in producing image recognition artificial intelligence, augmented reality, and video data. Alchera has 3D face analysis technology that can operate on smartphones, hand position recognition technology, as well as face recognition, behavior recognition, and object recognition technologies. We also developed unmanned facility monitoring technology. Through this work we have received three rounds of investment with our seed funding coming from Naver and series B recently led by Shinhan Capital. In addition we have attracted business from Naver, Samsung Electronics, SKT, LG U+,KT, and 16 other companies, research institutes, and public organizations as customers.
Q. What was your journey to starting a business?
Young Kyoo Hwang: The two of us have long developed face recognition technology at the Samsung Advanced Institute of Technology. We have experience in applying our technology to many products of Samsung Electronics and commercializing them. However, research subjects are limited when working at a large company. There was a common goal to do something innovative in a freer and open environment. At that time, artificial intelligence companies were starting to emerge, and we were confident that it was a field we were good at. Since the nation’s artificial intelligence technologies were only being released as products for large companies, the idea was to lead the artificial intelligence sector independently.
JBK: Hwang later moved to SKT’s Institute of Future Technology, where he was to be a technical advisor for the “SNOW Camera” technology, a popular beauty and makeup camera app. Snow recommended that we start a business because they needed 3D face analysis technology. In fact, we attempted to do a startup before, but failed several times. We continued to harbor dreams of starting a business and finally, we started our own because of Snow. Based on the technology we had developed, we could quickly commercialize the necessary technology for SNOW, so we immediately started Alchera. Usually, venture companies find the market after they create the technology, but we did it in reverse, we had to create the technology with the market secured.
Q. What kind of technology was needed in SNOW?
YKH: SNOW Inc. is a subsidiary of Naver. At that time, only 2D stickers were provided without face recognition technology. If you analyze your face with 2D technology, it’s less fun or interesting because it doesn’t have a three-dimensional, immersive feel. For example, when you put a regular cat ear-shaped sticker on your face and turn your head, the cat ear sticker doesn’t follow your face. However, by analyzing the face in 3D, the cat’s ears are shaped for your face because it has deeper information, which makes it possible for the cat’s ears to really stick on your head in a more three-dimensional and realistic way. Because foreign companies were applying 3D stickers in camera apps at the time, SNOW needed face recognition technology to implement 3D data in a 2D camera environment. Even Naver did not have its own technology ready at the time.
Q. What technology specifically has been applied to SNOW?
YKH: To be exact, facial analysis technology has been applied to Snow. To use more technical terms, ‘3D facial landmark tracking’ technology is applied. The camera app requires three-dimensional information about the eyes, nose, mouth, and jawline to attach a three-dimensional virtual object to the face. These virtual objects include hats, glasses, rabbit ears, and beards – whatever SNOW would like. It offers services such as 3D face stickers, virtual color lenses, hand stickers, and 3D animations. In particular, the 3D animation service analyzes the facial muscles of the user to create an animal emoticon or ‘animoji’, which can usually only be used with an iPhone X or camera with a 3D sensor. however, Alchera’s technology enables 3D analysis through software even with a 2D camera.
Q. Alchera has already done a lot of work. What are your main achievements?
JBK: Alchera has commercialized as many as four products in three years. This is what we are most proud of. In addition to SNOW Camera, we have helped bring LG U+’s AR IPTV “Kids Land 2.0,” an AR kiosk for photo-tickets being installed at Multiplex CGV, and Seongnam City’s smart security cameras to market. Kids Land 2.0 is a service that recognizes the faces and hands of children and allows them to experience AR at home which is used for education and entertainment. The AR kiosk or ‘My Poster’ is a smart kiosk that allows users to synthesize themselves into their favorite movies on movie posters through face and hand recognition technology in real-time, and print out or share the results through a photo-ticket or on social media. Seongnam’s smart security camera business is applied with a monitoring system that detects people, cars, and fire based on deep learning image recognition. We call it Visual Anomaly Detection or VADT and this is one use case. In addition to detecting passing cars and people, officials can search for a person’s age and description through Alchera’s technology.
Q. How did you raise the funds needed to develop this technology?
JBK: The funds needed to develop the technology were raised via investment as well as sales. Upon the establishment of the company, we received investment from Naver as I mentioned. We developed and supplied the technology in just three months so we guessed this impressed them. After that, we have received investments from Intervest and Intervalue in 2018, and more recently attracted investments from Soo Investment and Shinhan Capital in 2019.
Q. What is the secret of attracting investment?
YKH: Naver invested 1.5 billion won (roughly 1.3 million USD) in us. There has never been a case in which they have invested so much in a startup. In fact, unlike other startups, this was possible because our technological maturity was based on their own experience. We were able to attract two more investments after Naver because of the evidence that we have potential to supply and scale technology to large conglomerates. To display the potential for business, technological maturity seems essential, and the ability to supply technologies tailored to customers also seems important. If we continue that way and show investors the market’s growth potential, the investment appears to come naturally.
Q. Can Alchera’s sales be disclosed?
JBK: Sure. We earned 70 million won (60,000 USD) in 2016, 700 million won (600,000 USD) in 2017, and 2 billion won (1.2 million USD) in 2018. We’re targeting for more than 5 billion won or 4.5 million USD this year. We are considering listing on the KOSDAQ around the second half of next year.
Q. I don’t think it’s going to be easy to make enough sales to think about a listing in three years. How do you run your business?
YKH: Alchera has always agonized over the market, being most concerned about providing the right technology for our customers. It has not been difficult because we have long experienced supplying technology to large business units and getting them to release the products they want. You could say our strategy to continue to expand sales is by making use of our first sales.
Q. Alchera was honored by the Ministry of Science, Technology, Information, and Communication. What was so highly regarded?
YKH: Alchera succeeded in commercializing technologies in a short period of time after its founding, generating sales and contributing to the artificial intelligence industrialization of our country. In addition, we were highly regarded for our active employment of our young and talented staff contributing to job creation.
Q. How many people are currently working in Alchera?
JBK: Forty-eight full-time workers and 60 part-time workers now, but we are growing. The full-time job consists of 32 researchers and 16 operational and support workers, and among the researchers, about 18 hold either a master’s or doctorate. All part-time workers are video-big data workers. Image recognition technology requires millions of pieces of data by nature. Because the process of making data is relatively simple, such as taking pictures in person, manpower to make the video data is also used at our data centers in Vietnam.
Q. What areas do you plan to apply the technology to in the future?
YKH: At present, we mainly focus on facial analysis, and analyzing the rest of the human body. We will further upgrade our technology to analyze people in video for innovative use cases. We are currently working on unmanned facility monitoring technology which can monitor a facility 24 hours a day, 365 days per year without the need for humans to be put in danger, such as at electric utility substations and transmission towers. The technology is also expected to be used in safety areas such as fire detection.
Q. What kind of company do you want to build?
JBK: The company’s motto is to realize our dreams with AI. Since our establishment, we have been aiming to create Korea’s top artificial intelligence company, a globally competitive artificial intelligence company. We want to create a company that can bring people joy, safety, and happiness as the nation’s artificial intelligence leader.