Google expands AI virtual try-on technology to clothing, overcoming technical challenges | Generative ai certification google | Google machine learning course | Generative ai financial services | Turtles AI

Google recently expanded its AI-based virtual testing tool to include clothing, addressing the technical challenges associated with the complexity of this type of garment. Using a new training strategy and an advanced technique called VTO-UNet Diffusion Transformer, Google aims to provide an accurate and detailed virtual try-on experience. The expansion includes thousands of garments from brands such as Boden, Maje, Sandro and others, seeking to outperform competition from companies such as Adobe, Amazon and Walmart.

Key points:

•  Expansion of Google’s virtual try-on tool to include clothes.
•  Development of the VTO-UNet Diffusion Transformer technique to improve details and fit.
•  Involvement of hundreds of brands in the new feature.
•  Challenging competition from companies such as Adobe, Amazon and Walmart.

Google has expanded its AI-based virtual try-on tool to now include clothes, a category of clothing that is particularly popular among users. So far, Google’s system had already been successful in the virtual simulation of tops and blouses, thanks to technology that can realistically replicate the fall, creases and shadows of fabrics on virtual models in various poses. However, the greater complexity of the clothes, with their intricate details and variety of lengths and cuts, required the technology to evolve.

The initial method used by Google, based on diffusion techniques, encountered difficulties in accurately representing the detailed prints typical of dresses, such as floral or geometric patterns, due to the limited resolution of the images. To address these issues, Google has developed a new approach that starts with low-resolution images and gradually increases visual quality to high levels while keeping essential details intact. This improvement allows users to virtually try on a range of clothes, ranging from mini to maxi, without compromising the accuracy of the representation of the body and the garment itself.

A further technological advance is the introduction of the VTO-UNet Diffusion Transformer (VTO-UDiT), a technique designed to solve the problem of blurring body details when a garment covers a large part of the body. VTO-UDiT works to preserve the distinctive features of the person wearing the dress, providing a more accurate view of both the garment and the human figure.

This tool comes amid growing competition in the virtual try-on sector, with giants such as Adobe, Amazon and Walmart already launching similar platforms, offering customers the chance to virtually try on various garments. With this expansion, Google seeks to establish a dominant position in the market by offering a feature that aims to be superior to those of its competitors.

 The expansion of Google’s virtual try-on tool represents a significant step toward greater accuracy and variety in online shopping, solving technical difficulties related to the representation of clothes and improving the user experience.