Categories
Ξ TREND

New Google Shopping feature uses AI for pattern fitting


The service is currently available in the United States of America.

It will now be easier to buy from Google’s online shopping service. In order to avoid disappointment upon receipt of the ordered clothes, Google has implemented a feature. The latter allows the buyer to carry out a virtual fitting of the garment before confirming the order. The fitting is done using the AI ​​which models the outfit according to the indications given on a control body. This allows the buyer to have an idea of ​​how to look in the outfit.

A service first developed for women’s clothing

Google has announced that its virtual try-on feature is functional in the United States of America. The function that is operational from the AI ​​​​allows you to see the rendering of a garment on different body types. The user can specify all the details that can help him see a realistic preview of the fitting. This is possible thanks to the various human models available. There are different skin tones, sizes ranging from XXS to 4XL or hair types. Body shape can also be specified.

This innovative service still has its limits. It is only available for women’s items. Men will still have to wait. Another limit, only the top can be tried on and this within the limits of the brands taken into account. If the user wants to buy Loft, Everlane or H&M, he can fully enjoy this unique experience.

Before the end of the year, other brands should be integrated into the service.

With this feature, Google will be able to lower the figures of 59% of online shoppers who are disappointed with the purchase of an item of clothing because they expected it to look different on their body. Similarly, the 42% who do not feel represented by online clothing models will be able to benefit from better service.

To get an idea of ​​the level of realism of Google Shopping’s virtual fitting service, you have to look at the AI ​​that is used.

It is a generative AI model based on dynamic broadcasting. This makes it possible to have a realistic image with the way the garment wrinkles, bends or squeezes the virtual body designated by the user.