August 29, 2022
AI for Healthy Skin
MIS-UNIQ System Delivers Personalized Healthy Skin Management
In the present society, rapidly developing artificial intelligence (AI) has brought new ways of managing skins. For those who care about their own skin conditions, good skin care is very important. However, caring for skin is not an easy task because the influence of factors such as climate, health status and lifestyle, needs to be considered. A technology based on reliable data sources is expected to accurately analyze various skin conditions and hence contribute to providing effective skin care management.
MIS-UNIQ System, Understanding the Skin with AI
On November 18, 2021, the 2nd East Asia Skin Health Research Center Seminar was held in Shanghai, hosted by Yunnan Baiyao, a well-known large pharmaceutical group in China. At the meeting, "MIS-UNIQ", the customizable AI skin diagnosis and treatment system was officially released. The system was created jointly by OMRON and Yunnan Baiyao. OMRON has stepped into the field of intelligent skin management with its core technology, Sensing & Control +Think.
The system collects skin images of users for analysis. Based on the analysis models created by dermatologists, it is capable of identifying 86 different skin issues, allowing the users to accurately monitor changes in their skin and develop a subsequent treatment plan.
Image Collection, First Step to Understand the Skin
For the development of the MIS-UNIQ, the first step was to take photos so as to accurately capture skin characteristics, since they provide a basis for the analysis of skin conditions.
Cameras and light sources are crucial in accurately capturing skin data. The OMRON's Software Development Center team installed the MIS-UNIQ with a professional camera which is capable of capturing high-quality images thanks to the resolution of 32.3 million effective pixels and accurately presenting deep behind the skin.
Simultaneously, the R&D team configured light sources in varying wavelengths which helped the camera obtain images more clearly and accurately. This led to a new imaging technique called transdermal optical imaging, which allows oil detection and shows conditions deep beneath the skin.
Evolution in Light Source Testing
AI Algorithm to Analysis Deep Beneath the Skin
When capturing images, the MIS-UNIQ takes 15 pictures from three different angles for each user. The AI algorithm compiles the images to diagnose facial skin symptoms and generate pictures and analysis reports, which are instrumental to really understand the skin.
The OMRON's R&D team developed an image generation algorithm which creates an accurate facial map showing the areas of redness, brown spots and oil. At the same time, the team used the OMRON's facial recognition algorithm OKAO capable of detecting subtle facial expressions such as blinking and smiling in order to avoid those changes affecting the images. The facial maps serve as important indicators for providing tips for follow-up skin care and treatment plan.
By continuously evolving the algorithm, the accuracy of image composition improved and the time it took was greatly reduced. "In the early days, it took us around five minutes to capture 15 images. The time has now been shortened to one and a half minutes through the improvements of algorithm, equipment performance and capture process, greatly improving the user experience", said Zu-Biao Fu, who is a member of the OMRON Development Team, and in charge of algorithm development.
Zubiao Fu, Member of the OMRON Development Team
Story Behind the Development
The biggest challenge the development team faced was to accurately identify and extract facial features. To tackle this problem, the OMRON team studied more than 1,000 samples of people's faces.
"If two devices were used to take photos, we wondered if they would project the same feature points. What about the images of the same sample taken by the same device? Would they be the same? We also wondered whether the results would be affected if the images of the same sample were taken by the same device but in different locations, and so we tested repeatedly." recalled Wang Dong, the Head of the Solutions Department of OMRON China Software Development Center.
Dong Wong standing beside MIS UNIQ Skin Tester
Samples were collected from different age, gender and region, even included photos of the same person taken in different times and regions. At the end of the extensive research, the OMRON team finally completed the AI algorithm enabling the system to accurately distinguish the underlying condition of the skin according to the photos, and generate high precision images.
Possibilities spread into the Future
The MIS-UNIQ system capable of detecting skin conditions anytime and anywhere may become available for home use in the future. Moreover, if the OMRON's leading environmental sensors are combined with the skin management device, a more comprehensive solution for skin management based on environmental information and data could be provided to users.
Under the whole wave of technologies transforming life, OMRON will continuously challenge itself to creating a better life and be a driving force for innovation.