Shopping for the latest fashion trends online can be both exciting and challenging. One of the main hurdles is visualizing how a new style will look on you, which often results in a cumbersome returns process for both shoppers and retailers. To address this issue, Google has expanded its virtual try-on (VTO) feature to include dresses, enhancing the shopping experience with the power of generative AI.

Elevating the Virtual Try-On Experience

Last year, Google introduced a virtual try-on tool for women’s and men’s tops, allowing shoppers to see clothing details such as shadows, wrinkles, and draping on models. This innovation proved beneficial, with virtual try-on images receiving 60% more high-quality views compared to traditional shopping images. Users engaged more with these interactive visuals, and the likelihood of visiting a brand’s site increased.

Building on this success, Google is now expanding this feature to one of the most searched apparel categories: dresses. This new addition allows shoppers to visualize how various dress styles will look on models who reflect their own body type, helping them make more informed purchasing decisions. Shoppers can explore dresses from hundreds of brands within Google’s Shopping Graph, including notable names like SIMKHAI, Boden, Staud, Sandro, and Maje.

How It Works

For shoppers in the U.S., using the virtual try-on feature is straightforward. When searching for dresses on Google Search, look for items marked with a “try on” badge. Clicking on these styles will bring up images of the dress on a diverse range of real models, available in sizes from XXS to XXXL. By selecting a model that matches your own body type or preferred style, you can get a more accurate sense of how the dress will look on you. Once you’ve found the perfect dress, simply click through to the retailer’s site to make your purchase.

Behind the Technology

The virtual try-on feature leverages cutting-edge generative AI technology developed specifically for this purpose. Google’s VTO tool uses a diffusion technique, which generates high-quality, realistic images from scratch. However, applying this technology to dresses posed unique challenges. Dresses are more detailed than tops, with complex elements like draping, silhouettes, and prints that are difficult to capture accurately.

To address these challenges, Google’s research team developed a “progressive training strategy” for VTO. This approach starts with lower-resolution images and gradually improves the resolution during the diffusion process, ensuring that fine details like pleats and prints are rendered clearly. Additionally, Google introduced the VTO-UNet Diffusion Transformer (VTO-UDiT), a new technique that preserves a person’s key features while seamlessly integrating the dress. This prevents “identity loss” and maintains accurate portrayals of both the dress and the wearer.

Impact on Marketers

For marketers, the expansion of Google’s virtual try-on feature offers significant advantages. It enhances customer engagement by providing an interactive and personalized shopping experience, leading to increased time spent on product pages and higher likelihoods of conversion. The ability to visualize dresses on models of various body types also aligns with inclusive marketing strategies, catering to a diverse audience.

Moreover, integrating virtual try-on into the shopping experience can drive more traffic to brand websites, as shoppers are more inclined to visit and purchase from brands that offer this innovative feature. Marketers can leverage this tool to differentiate their offerings and improve customer satisfaction, ultimately boosting sales and brand loyalty.

Google’s new virtual try-on feature for dresses represents a major advancement in online shopping, making it easier for consumers to envision how clothing will look on them without the hassle of returns. By incorporating cutting-edge AI technology, Google not only enhances the shopping experience but also provides marketers with valuable tools to drive engagement and conversions. As this feature continues to evolve, it promises to redefine how we shop for fashion online.