'Smile can be discounted', 'took it away, go and pay' ... In the fourth World Internet Conference - Wuzhen Summit, Lynx unmanned supermarkets as the only independent exhibitors unveiled the project. Cat no one supermarket not only no one to pay, also introduced the 'Happy purchase' emotional marketing, 'smiling at the goods can be discounted', attracted the scene guests eager.
On December 3, the Fourth World IoT Congress was held in Wuzhen, Zhejiang Province. At this conference, more than 400 representatives of the world's leading Internet and technology innovation enterprises were brought together to discuss emerging emerging issues such as the Internet, Cloud Computing, Big Data, Internet of Things, and Artificial Intelligence Technology to explore the big guys gathered in this session, black technology will not be absent.
This does not, Jack Ma unmanned supermarket as the only independent exhibition project debut this session, Ali may think no one supermarket technology is not enough 'black', so the 'emotional marketing' debut.
It is reported that Lynx unmanned supermarket use 'behavioral trajectory analysis', 'emotional recognition' and 'eye tracking' technology, which launched the first emotional marketing 'Happy Buy' when consumers stare at a product, the system will be timely Capture the expression of the emotion and judge the emotion of the consumer according to the expression of the emotion, calculate the degree of preference of the consumer for the commodity, and then give different discount according to the degree of preference, and different consumers will get different discounts on different emotions expression of a commodity, This is Ali's emotional marketing 'Happy Buy'.
Israeli Nobel laureate Danielle Sheikh Thamman and Chinese Academy of Engineering Ni Guangnan gave praise after the experience, so emotional marketing in the end what a ghost?
Relationship between emotion and shopping
At present, the upgrading of consumption has almost become a stub of the retail industry. What is the escalation of the so-called consumption upgrade? The demand for quality is on the rise; other attention such as personalization and product added value are given more attention. These ultimate goals are to give Consumers bring more pleasure experience, in fact, this is the kind of customer satisfaction achieved.
I do not know if you are not aware of it, the movie often shows such scenes: the actress robs the heart and uses crazy shopping to heal the wounds, and this is a very typical example of emotions and shopping.Human itself is a very emotional animal, and these emotions are closely related to shopping , You are timid, so buy self-defense appliances; you are afraid to be looked down upon (vanity), so close to the name brand shopping to some extent to meet some of the emotional loss of consumers, although the completion of such a lack of shopping is only a Hallucinations, but shopping does create the illusion of being lost.
It is precisely for this reason that emotional satisfaction and building are of crucial importance in marketing activities.Many kinds of emotions are very diverse, and American psychologist Paul Ekman classifies the basic emotions of mankind into six categories: happiness, sadness, anger, disgust, Surprise and fear.Marketing is not the focus of attention on how many emotions human beings, but whether this emotion can arouse the resonance of consumers, resonance after the action is the key to marketing, so happiness, fear, disgust, anger four The kind of 'arousal' emotions are widely used in marketing.
Pictured above is a picture of the packaging of cigarettes in Hong Kong. This kind of public service advertisement uses exactly the fear of human beings. Its horrifying picture is believed to deter many non-smoking friends. As for the use of happy emotions, everyone just needs Open the major TV stations and video sites, full screen is no reason for the happy atmosphere.
How does the machine recognize human emotions?
Emotions are actually a combination of people's feelings, thoughts and behaviors.When talking about how machines recognize human emotions, it is necessary to talk about how human beings recognize emotions.At the beginning of human beings, human emotions were very naive In the process of growing up, the ability to perceive emotions becomes more and more acute and accurate by continuously interacting with others. That is to say, human beings acquire this sensitivity from very limited experience training, which is highly summarized One of the strengths of humankind is its ability.
AI is anthropomorphic, that is to say its recognition of human emotions is the same as that of human beings. However, it does not possess the highly summarized ability of human beings. It relies on huge data exercises and relies on logic and rules And training to acquire this ability, its accuracy and data practice is basically proportional to.
And human rely on facial expressions, sounds and movements to identify different emotions, the machine to recognize human emotions can rely on more ways, in addition to expressions, sounds and movements of these non-physiological signal recognition, the machine can also use the device to detect human heart rate , Skin impedance, respiration and other physiological signals to identify the corresponding emotional state, while the more ways to detect, the higher the accuracy of emotional recognition.
Of course, how to use more testing equipment to identify human emotions is not the focus of this article. This article will focus on how machines judge human emotions by recognizing human face images.
Before the recognition of emotion, the first thing the machine needs to judge is that it is a human face, not the face of other species. Due to the high complexity of non-human faces, the detection needs to consider the size, rotation, posture, expression, Occlusion, light changes and many other factors, so to determine the face and non-face generally use the pattern recognition method (can be simply understood as the exclusion method), the two different modes of classification, and then the image is placed on the face image database Conduct training and testing to determine.
In determining the face, the next is based on facial expressions to identify human emotions.Here we must introduce the character here is the aforementioned American psychologist Paul Ekman, Ekman through long-term observation of the original tribal groups and modern people's facial expression found, Human facial expression has commonality, so Ekman divides it into a number of independent and interrelated units of motion (AU) according to facial anatomy, and analyzes the movement characteristics of the units and the main areas they control As well as the related expression, and thus the standard movement of facial expressions, is the famous Facial Expression Coding System (FACS), which has an accuracy of 90% for facial expression capture.
Facial expression coding system (FACS) provides theoretical support for the machine to judge human emotions through face recognition, and then the machine digitizes the relevant information of the face image.First, after the machine recognizes the human face, it extracts the key muscle points of the face, The function of these key points is to mark the facial features and contours of human beings in order to locate the key positions of the human face (eyes, mouth, nose, etc.). The more the key points of the machine are extracted, the higher the recognition precision is. The corresponding performance requirements are also higher (at present the extraction of 500 key points are more used).
The different facial expressions of people are essentially the changes in their position (relative position, absolute position, distance, angle, etc.) and then comparing these changes to the already digitized Face Expression Encoding System (FACS). The human equivalent Emotions will be presented. It should be noted that the 43 muscles on the human face can combine more than 10,000 expressions, but not all emotions have an emotional meaning (about 3000 kinds of expressions), but are currently being applied in marketing Most of the six basic emotions mentioned earlier (happiness, sadness, anger, disgust, astonishment and fear), especially happiness, fear, disgust, anger, four kinds of 'arousal' mood is more widely used!
Now talk about the 'Emotional Marketing' pushed by Lynx unmanned stores that debuted at the World Internet Conference. From the introduction, we can see that when a user gazes at a product in an unmanned supermarket, the system captures their emotions in time , And according to the emotional range, quickly calculate the degree of consumer preferences for the goods, give different discounts.Obviously, Lynx is through the consumer face image recognition to determine when consumers face the emotions of the commodity, and then Given the appropriate discount, and this is the author in the previous introduction of the principle of practical application, Lynx will be landing to some extent, showing a future retail development trend!
Finally, I also have to worry about the development of this technology will be to consumers 'opposite', that is, the use of emotional identification to enhance the price of goods in disguise (the specific method is no longer detailed, the reader is not hard to imagine), in fact, A similar event has taken place. A website platform uses technology to learn about the spending of a particular customer and then customize 'different' prices.