A new patent granted to Samsung could result in shopping recommendations fueled by the emotional response potential buyers show while looking at products. Filed with the World Intellectual Property Office under the designation ‘WO2018143630,’ the new patent covers both the device and deep learning-based methodology required to make that happen. Since the resulting system is powered by A.I., machine learning is central to the process. That would be utilized to effectively teach Samsung’s system to recognize various facial expressions. It would then gauge the reaction of a user observing a “displayed product” in order to determine whether or not that individual is happy with what they are looking at. In fact, the reactions would range across at least six emotions according to the diagrams on the patent. Those are documented as including anger, sadness, surprise, joy, happiness, and more. Finally, the system would pull together a “plurality of product sets” to be recommended starting from that as the basis.
The system would also go further to factor in an assessment of weight, gender, and age as well as previous purchase history and preferences. In practice, it’s difficult to say whether Samsung plans to use the technology in one of its own smartphones or in something else. The cameras are shown capturing the entire person in Samsung’s diagrams, for example. So it wouldn’t necessarily be out of the question for a retail location to embed the system into its displays or in-store mirror so that a person can get recommendations they’ll enjoy without wandering too much. That solution or something similar to it would also make sense since the system is shown and described as offering not just one or more items of the same type. Instead, it recommends entire sets of items such as an entire outfit and accessories that go well with the outfit but still fall within the recommendation parameters.
Of course, machine learning means that the recommendations should also be able to sync up with an individual user profile and learn their habits over time. Similar technologies, including an Amazon product that applies recommendations based on opinions of fashion experts. With that said, adding facial expressions to the mix would effectively help the artificial intelligence get a better grasp on whether or not the person in question will actually enjoy wearing the outfit. That could be a fairly big leap forward for fashion-aware people regardless of what kind of final format it ends up in.
