The information that artificial intelligence system needs to perceive mainly comes from its vision, hearing and touch, and smell is also a valuable direction. At present, researchers have been studying odor sensors and their applications in environmental protection and other fields.
During my stay in China International Import Expo(CIIE), in addition to "hard-core" intelligent technologies such as bionic robots and automated production lines, AI beauty consultants, virtual makeup artists and customized skin care products driven by artificial intelligence also attracted many people’s attention. In recent years, artificial intelligence, which shines brilliantly in the fields of image, video and speech recognition, has gradually penetrated into the beauty industry. The market shows that the perfume prepared by artificial intelligence has received good feedback from consumers.
"Perfume preparation is a complex system engineering, involving raw materials, processes, users, markets and many other factors. Searching for the right solution in the vast parameter space is a huge workload. It is difficult to exhaust all feasible solutions only by relying on manual experience, and it is likely to miss the best solution that is most popular with the market and users." Sun Zhenan, a researcher at the Institute of Automation, China Academy of Sciences, said in an interview with a reporter from Science and Technology Daily.
AI flavoring does not depend on fragrance to judge the order of putting spices.
"The information that the artificial intelligence system needs to perceive mainly comes from its vision, hearing and touch, and the sense of smell is also a valuable direction. At present, researchers have been studying odor sensors and their applications in environmental protection and other fields." Sun Zhenan said.
Previously, IBM made an artificial intelligence perfume apprentice — — Philyra。 Philyra is a creative artificial intelligence, which can study the existing perfume formulas and compare and analyze their components to develop new perfume formulas. Philyra can also obtain the popularity of perfume in different genders, ages and best-selling places. Combined with big data algorithm, Philyra will output a new perfume formula, which can achieve good results in the predetermined target population.
Sun Zhenan explained that unlike perfumers who need decades to train their sense of smell, artificial intelligence does not rely on the sense of smell to make perfume. AI Blending uses advanced machine learning algorithms to analyze and learn information such as perfume formula, raw materials, historical sales data and industry trends to predict human preferences, thus creating a new perfume formula for the target population.
David Appel, a master perfumer, did an experiment with Philyra: a perfume made entirely by Philyra was model A; The b model is mainly made by Philyra, and the perfumer is used as an auxiliary modification; The C model is led by the perfumer and assisted by Philyra. After the completion of the three perfumes, they were put into test. The test results showed that most people chose the A perfume created by artificial intelligence.
"Compared with artificial perfuming, AI perfuming does not rely on fragrance to judge the order of perfuming, but uses deep learning algorithm to analyze and customize the blending method, which is a method different from traditional perfuming thinking; AI perfume will not be influenced by personal preference, experience and cultural prejudice, but digitally quantify the sense of smell and explore more creative composition combinations and formula; With its powerful computing power, AI perfume can quickly calculate the most popular perfume formula for target customers, shortening the perfume development cycle. " Sun Zhenan pointed out.
It is also used in intelligent medical beauty, hairstyle design and other scenes.
According to Euromonitor Consulting’s forecast, 2019— In 2023, China’s cosmetics industry will maintain a compound growth rate of 8.3%. Faced with the growing cosmetics consumer market and the popularization of concepts such as scientific skin care and functional skin care, major beauty and skin care businesses are also actively using cutting-edge technologies such as artificial intelligence to launch customized products that better meet consumers’ individual needs and bring more comprehensive product experience.
Sun Zhenan said that in addition to preparing perfume, artificial intelligence is also used in many scenarios such as intelligent medical beauty, hair design, virtual makeup, customized skin care products, skin diagnosis and so on.
Sun Zhenan, a researcher from the Institute of Automation, China Academy of Sciences, and his team of researchers have explored the editing of face images and the application of beauty, medical beauty and entertainment.
For example, at the IEEE International Conference on Computer and Pattern Recognition in 2019, they proposed a face age transformation technology based on wavelet domain, which can predict the change of face with age, and this technology can also be applied to personalized recommendation of skin care products; In 2020, they proposed a face attribute editing technology based on sample importance sampling, which can edit the skin color, hair color, adding or removing glasses and other attributes of the face, and can be applied to the makeup design of the beauty salon industry to provide users with diverse choices and references; In 2020, a face composition editing technology based on reference images was proposed, which can manipulate the shape of facial features and assist medical cosmetology, give more guidance to doctors and patients through the preoperative prediction of plastic surgery and trauma repair surgery, and make a more scientific cosmetic plan.
"The application technology of artificial intelligence in the beauty industry has saved labor costs for enterprises, improved customer service efficiency and brought convenient and comprehensive user experience for consumers to purchase products." Sun Zhenan said.
Sun Zhenan pointed out that from the current application, the personalized customized skin care system allows each user to have a dedicated skin care beautician, analyze the photos provided by the user by using face recognition and face analysis technology, and evaluate their skin condition, and then use the recommendation algorithm to select the most suitable product from the existing products for recommendation; The facial beauty evaluation system provides valuable methods and tools for medical cosmetology, which can assist the hospital in cosmetic surgery. By comparing the detected data with the standard beauty model, it provides the cosmetic surgeon with the best scheme for improving each part. AR intelligent makeup trial uses face recognition and virtual makeup rendering algorithm to realize the makeup effect of products with different brands and colors on the face, which reduces the cost of makeup trial, expands the beauty retail terminal and meets the demand of fast fashion.
For aesthetic standards, people and machines will influence each other.
Different people have different opinions about beauty, and it is difficult to quantify and form a standard. However, "everyone has a love for beauty", and the aesthetic concept of the public still has rules to follow. For example, a beauty who is truly beautiful can still be recognized by almost everyone.
"Artificial intelligence’s understanding of beauty requires the training of big data. At present, the definition of beautiful faces and the face value of artificial intelligence cannot be separated from the supervised data marks of human beings, because artificial intelligence has no understanding of beauty itself, and its grasp of beauty depends on the face data it has learned and the label data set that artificially defines and quantifies beauty." Sun Zhenan said.
David Appel thinks that artificial intelligence is not a threat but an object that needs cooperation. Machines have no unnecessary feelings and prejudices, which can significantly enhance human creativity and provide the most novel algorithms.
Previously, the CARTO artificial intelligence fragrance blending system was produced in the well-known perfume company Givaudan. The system used the odor atlas collected by IBM, hoping that the fragrance could maximize its olfactory performance in the formula.
For another example, Shiseido Optune personalized customized skin care system consists of a smartphone application Optune App and a dedicated machine Optune zero. Optune App displays the user’s skin condition (skin texture, pores, moisture content, etc.) by analyzing the photos taken by users with machine learning technology, and sends the data to Optune Zero through the cloud, which selects the most suitable combination of essence and moisturizing products from existing products according to the algorithm and produces them directly from the machine.
"In order to eliminate the preference of artificial intelligence for human aesthetics, the data used in the training algorithm should be as comprehensive and diverse as possible, and multi-dimensional and multi-angle standards should be integrated for learning. However, it does not rule out that autonomous evolutionary intelligence has the ability to discover beautiful people and things unsupervised. For example, machines may automatically search for star photos that attract public attention to learn aesthetic standards independently. " Sun Zhenan said.
In Sun Zhenan’s view, with the application of artificial intelligence in the beauty industry, people and machines will influence each other on aesthetic standards. "On the one hand, human beings have a personal preference for the standard of beauty in a broad sense. When algorithm experts use a certain definition of beauty as a standard for machines to learn, machines tend to be aesthetic standards of this definition. On the other hand, when the machine recommends the medical beauty scheme it thinks for a long time, the user aesthetic system of the artificial intelligence system will also be affected. "