Youtube wants to be able to recognize products shown in its videos to be able to recommend videos. It would also offer the option to buy them through links on the page.
The YouTube portal has started testing a new content detection system to recommend videos to users. This time it is a system that would detect the products that we see in the videos. After detecting them, he will recommend videos related to him in case we are curious.
No products found.
A new way of recommending videos: for what it contains
This idea has come from the idea that if we see something on the screen, we will be curious to see if the piece of hardware that a youtuber uses is good or if the game they are playing is promising, we will look for a video that makes an assessment of it . For now, it has begun its journey in the United States. It is not the first time that YouTube wants a system that recognizes products what is seen on the screen to recommend videos.
At the time, this neural network for detecting YouTube products recommended links to stores that had said product for sale. Now his priority is to detect them to recommend other videos of the same one on the platform. At the moment, it puts us under the mobile player a list of products that have been detected within the videos we are watching.
It is not known if this algorithm will be released outside the United States, because it is at a somewhat early stage to be extended to other countries. If it goes ahead, YouTube could get more revenue from the platform by taking out many product sales links of all kinds that are intentionally or accidentally showing the content creators we are seeing. It would also greatly improve your video recommendation algorithm, as it will be able to predict whether a user is interested in something shown in a video and would like more information about it. But this will work as long as you find the necessary video to answer your question.
Source: 6to5Google
[irp]
