Google RealFill, an AI-based model for completing generated images, spotted on the trademark list

Google RealFill could be the tech giant’s latest bid to innovate artificial intelligence (AI)-powered image generation in a user-centric application. Recently, a research paper and website named RealFill was spotted online that performs image completion and painting based on reference images and creates a target image. The company also appears to have filed for trademarks for its designed logo for the AI ​​model-based product. It should be noted that the new AI model uses computer vision and pattern recognition algorithms and is trained using random masking techniques.

Github page and preprint paper of the AI ​​model were noticed recently from Android Authority. The publication also found trademark applications filed under the name Google LLC with the US Patent and Trade Office (USPTO) and the European Union Intellectual Property Office (EUIPO) announced. Based on these, it appears that the tech giant has not only reached the end of the research phase for the AI ​​model, but also has plans to introduce it as a commercial product.

According to the GitHub page, RealFill is described as “a new generative image completion approach that fills missing regions of an image with the content that should be there.” Essentially, the AI ​​model can scan multiple images of an object in the same setting and then use that reference to generate a preset image. As a tool, it can be used when a user has clicked several photos of a subject but failed to capture the perfect photo. AI can process these images and generate an image that doesn’t even exist.

RealFill is a generative AI model that uses computer vision to understand the object and environment of reference photos and can handle various aspects of it, including technical specifications such as sizes, colors and shapes, as well as contextual understanding of different objects. Using this information, it can then create a target image in a new plane and fill in details that may not be present in the reference images.

While it’s hard to predict Google’s plans with RealFill, last year the company shipped a feature in the Google Pixel 8 series called Best Take, which can process multiple photos in a group shot and let users choose the best expression from each image. , to create a final photo. Compared to RealFill, this feature looks very basic, but based on its application, the tech giant may be preparing the AI ​​model to make it a feature for Pixel phones.