After tinkering with AI for years, Google has finally begun to show us the practical results of this work. There was no better sign of this than atGoogle I/O 2023, where the new Pixel devices seemed to be an afterthought to all the talk about AI. Google clearly recognizes the power of AI in their technology, and today you can see what many people could consider the pinnacle of this technology; trying on clothes.
Anyone whose bought clothing online knows the danger. No matter how hard you try, there’s a good chance the final item won’t fit. Google’s newest AI feature generates a garment on models ranging from size XXS-4XL with various skin colors, using the Monk Skin Tone Scale as a guide. The idea is to give you a better idea of how a product will look on you before you buy it.
This limited availability is generally expected for new Google software features, but it claims it will expand this feature to more retailers while adding men’s clothing. It’s not a perfect tool, but themethodology behind this featureis pretty interesting.
To create this technology, Google first rounded up models with various body sizes and skin colors, using the Monk Skin Tone Scale as a guide (there’s no consideration for height or muscle tone yet). Google then applies diffusion to each model, using the picture of the model and the garment as inputs to generate the final images. The idea here is that there will be so many results you should be able to find a rough match for your figure.
To complement this new tool, Google has implemented AI-powered filters to help you find the perfect product. Once you’ve picked a garment, you can refine the results (e.g., by color or pattern) to see products across retailers.
While these features aren’t ascontroversialordramaticas other AI tools, their practical applications mean many will find them more relevant to their lives than most AI tools floating around. However, the virtual try-on tool is limited, currently only available for US shoppers looking for women’s clothes from select retailers like Anthropologie, Everlane, H&M, and LOFT.
