Recently, Google announced the expansion of its artificial intelligence (AI) tools to cover one of the most frequently searched clothing categories—dresses. The company has introduced an AI virtual try-on feature on the Google Shopping platform, allowing consumers to get a more accurate sense of how different dresses will look when worn, thereby enhancing the online shopping experience.

Last year, Google pioneered a virtual try-on tool for tops, using generative AI to help shoppers better visualize garment details. After implementation, Google found that the virtual try-on feature not only increased shopper engagement but also significantly boosted traffic to brand websites.

Now, with just a few clicks, consumers can browse hundreds of dress brands on Google Shopping, including well-known names like SIMKHAI, Boden, and Staud. U.S. consumers can view how dresses fit on various real-life models by entering "dresses" in the search box and clicking on styles marked with "Try-On." The sizes range from XXS to XXXL, allowing users to select the best fit. Once consumers find the desired style, they can click through to the website to make a purchase.

This feature is powered by Google's generative AI technology, specifically designed for virtual try-on (VTO). While it successfully generates high-quality, realistic images of tops and shirts on models, handling dresses posed two unique challenges. Firstly, the design of dresses is more complex and diverse, involving different drapes, contours, and patterns. To tackle this, Google's research team introduced a progressive training strategy that trains the AI from low-resolution images to high-resolution ones, allowing for better detail representation. This method ensures that every fold and print is vividly captured.

Secondly, because dresses cover more of the body, traditional "erase-and-replace" techniques could result in body features being obscured or lost. However, Google's latest technology—the VTO-UNet diffusion transformer—addresses this issue by preserving users' key features during garment replacement, ensuring the accuracy of the virtual try-on experience.

The launch of this new AI virtual try-on feature is set to enhance the shopping experience for consumers while providing brands with a better platform for showcasing their products.