GoogleThe company issued a press release on Thursday, September 5, announcing the expansion and enhancement AI Virtual Try-Ontool, which now allows users to virtually try on thousands of dresses from hundreds of brands including Boden, Maje, Sandro, Simkhai and Staud.
Google said that dresses are one of the most searched clothing categories in the tool, but because dresses are more delicate and complex than other clothing, existing AI diffusion technology has difficulty capturing print details such as floral or geometric patterns, which poses a great challenge.
Furthermore, since dresses usually cover large parts of the body and come in different lengths (e.g., midi, maxi, and mini), putting a virtual dress on a person often results in blurred body details.
After Google released its AI virtual try-on tool last year, it used its own diffusion technology to create high-quality, lifelike images of tops and shirts, which can simulate various postures of clothing on real people, such as drape, fold, attachment, wrinkles and shadows.
Google has come up with a VTO-UNet Diffusion Transformer (VTO-UDiT)The new technology erases and replaces the skirt while retaining the person's features, resulting in a more accurate depiction of the skirt and the person wearing it.
Virtual try-on technology is designed to take the guesswork out of finding the right clothing for customers of all body types. Many companies (Adobe, Amazon, and Walmart) have launched their own tools that allow customers to virtually try on a variety of clothing, including dresses. However, with this new expansion, Google seems to be looking to go one step further than its competitors.