Google Launches AI Virtual Try-On Feature, Allowing Shoppers to Visualize Outfits on Models

Google has launched an innovative AI-driven virtual try-on feature that enables online shoppers to see how clothing fits on various models. According to Senior Product Manager Lilian Rincon, users can simply click on items labeled 'Try On' in search results and select a model that resembles them. This feature helps prevent the disappointment of purchases that don't match a person's body type, skin tone, or hair color, ultimately minimizing the hassle of returns.

Shopping online presents the challenge of not being able to physically try on clothes. Typically, shoppers rely on model photos to envision how garments might look on them. However, if a shopper's appearance or body type differs significantly from the models', this visualization can often miss the mark. Google’s virtual try-on feature addresses this limitation by allowing users to choose models that reflect their own body shape, providing a clearer picture of how the clothing will look before making a purchase.

Powered by a generative AI technology called TryOnDiffusion, Google’s system analyzes clothing images and predicts how they would fit on models ranging from XXS to 4XL, showcasing a variety of poses. While Google is not alone in offering virtual try-on capabilities—Walmart also provides a similar feature—Google’s approach allows for greater pose variability rather than just a straightforward overlay of models facing forward.

Currently, Google’s virtual try-on feature is available for select women’s tops from brands like Anthropologie, Everlane, H&M, and LOFT, with plans to expand to more options and include men’s tops by late 2023. This advancement marks a significant step in enhancing the online shopping experience, making it more personalized and efficient for consumers.