Just as the human brain processes images, computers are trained to process images just like humans. Image recognition, which is now a hot high-tech, was born as early as 1966. The original idea of the inventor was to attach the camera to the computer, so that the computer could easily see what the camera actually saw. However, image recognition still needs human training.
Today, image recognition has become a part of our daily life and is sold by brands and retailers. Technology giants such as Apple, eBay, Facebook, IBM and even Kim Kardashian are promoting the development of image recognition. Google's advanced image search is constantly learning and developing. Apple and Facebook can identify a person by uploading images.
Facebook researchers and engineers have trained an image recognition network of up to 3.5 billion Instagram images with up to 17000 tags. But how do brands and marketers use this emerging technology?
Brands using image recognition
EBay recently launched ImageSearch, which allows users to take a photo or select a photo from the device and search for eBay items based on that image. This new feature uses machine learning to allow shoppers to use images when searching for matches. Compared with the previous text search, image search has taken a big step forward. I have to say that anything that makes it easier for consumers to shop is a good thing.
Craze's Screenshop is Kim Kardashian's image search application. All users need to do is upload a screenshot that you like, and then it can immediately purchase it on Screenshop. It not only shows what you want, but it also promotes sales by showing similar clothes and accessories and more affordable choices. This application has become a buyer for users, and it can be accessed anytime, anywhere, and allows users to browse at their own convenience.
What these two companies are doing is using relevance search as a means of personalized shopping experience.
Marketers using image recognition
Your friend sent a photo of the restaurant, but in fact, this photo contains more information than that. Basic elements such as location, tag friends, and tags provide some recognition features. However, image recognition analyzes everything in an image. It can judge whether you are drinking beer or vodka by processing the color of the wine in the glass, and even get the name of the restaurant from the menu on the table. The brand and fabric of clothing can also be identified.
This information provides many data points about us as consumers. Of course, it is also very valuable for marketers. If you are a KOL who only wear famous brand clothes and drink champagne, advertisers will capture this information and use it to attract more users. Marketers and technology companies are already using daily search information to provide related advertising. After all, we have all experienced the uncomfortable experience of searching for something on the Internet, and then it will be redirected everywhere with you.
The Road to the Future
Emotion analysis is a new research field in image recognition.
Facebook has several patents in this field, one of which is called "emotion detection and content delivery technology". This is to use the camera to track the emotional state, watch different things at the same time, such as dog videos, and serve future content, just read the user's emotional state. This type of image recognition adds a new dimension to personalization.
However, there are still many problems at this stage. There are many examples of image recognition to demonstrate these problems. For example, if the same person has different hair styles or colors, the computer will sometimes recognize them as two different people. In reality, if the computer is not smart enough to recognize a small change or nuance, then image recognition will be confused and it is actually a person.
In the past 50 years, image recognition has undergone tremendous changes. Facial recognition companies such as RetailDeep can use smart phone cameras to identify and record consumers when they enter the store, and salespeople can access their entire purchase records accordingly, easily combining online and offline, so that their sales can be more targeted in the future.
At the same time, devices such as Echo Play and Google Hub can even use the front camera to understand facial emotions in two-way conversations. Therefore, for marketers, image recognition is not an illusion, it is coming, and will also greatly affect the existing marketing methods.
