You can see the photo on Facebook, if you’re interested in an object that appears to it, and decide to buy it. But what and how? The Facebook has unveiled a tool for computer vision called GrokNet , on the basis of artificial intelligence, which identifies the goods in the pictures. From there, it is possible to generate a link which allow you to purchase the items you identified.
- Facebook and Instagram have leased the Stores to the companies that promote the products
- How to create your Avatar on Facebook
The GrokNet it is wide-ranging, being able to identify the parts of a garment, the decorative items are present in the picture. The tool has been designed to be able to take you straight to the item in the image, and hence, the factors such as color, texture, style, and up to the mark, may be taken into account.
According to Facebook, the GrokNet this is the first of a number of technologies based on artificial intelligence that will be available on their e-commerce platforms in the coming months. The novelty promises to improve the buying experience of the consumer, and in the process of selling on the retailer.
In this way, the consumer may be able to find exactly the item you want, it is not a similar as the original could not be found. Already, the seller will have a chance to display their products so that more of it occurs.
Thus, when a seller post a picture of a sofa in the Facebook Marketplace, for example, the GrokNet to identify the characteristics of a product (such as color, material and style) to make it easier to find for users who are looking for an item with the same attributes.
Another example is that a furniture store can show you a picture of a room, the GrokNet identify the products available there that are for sale, and you point the links to them. The user’s view of the objects in a real-world environment (or close to it), you’ll have an easier time deciding to buy the product.
This is what Facebook says it will. It is not yet clear whether the technology will actually have the level of precision, pointed out the company notes that the GrokNet is already able to recognize correctly 90% of the items in the photo to the Facebook Marketplace in the category Home & Garden for example.
In their favor, Facebook has a huge base of data is the algorithm used by the GrokNet they have been trained, with about 100 million images, with the majority originating in just the Marketplace, the environment in which the system is currently being implemented. This means that, if the technology is not as accurate as promised, it should at least come close to it in the near future.
This is just one of the efforts of Facebook to boost e-commerce on its platform. In addition to computer vision, the company aims to be complementary to these environments, augmented reality, 3D contents (to be able to view the product in 360 degrees), and digital assistants. (see the picture above).
With the information, The Verge, Engadget.