Microsoft is working on a wrinkle prediction tech that accurately depicts how individuals age

The tech can be used to create realistic images of faces.

Reading time icon 2 min. read


Readers help support Windows Report. We may get a commission if you buy through our links. Tooltip Icon

Read our disclosure page to find out how can you help Windows Report sustain the editorial team. Read more

Microsoft wrinkle prediction

Imagine a tool where you upload a picture of yourself, and moments later, the tool displays the image back to you, showcasing how your face will wrinkle. Microsoft is working on such a wrinkle prediction tool. On September 1, 2022, the company introduced a groundbreaking method for computing an image depicting a face with an expression and wrinkles, and the patented technology was made public a few days ago.

The technology is not your ordinary Snapchat or Instagram filter; it’s a sophisticated process involving a 3D polygon mesh model of a face, which showcases the wrinkled face in all sorts of expressions and emotions.

The tech starts with a tension map computed from the 3D polygon mesh model to generate your wrinkled face. This map is quite the brainiac; it knows exactly how much compression or expansion is needed to transform a vertex to model the face with a neutral expression or add more expression.Microsoft wrinkle prediction

But the real game-changer here is the use of wrinkle textures. These aren’t just textures; they are computed by aggregating maps of your face with different expressions using the tension map. A neutral texture shows the face in a calm state, a compressed wrinkle texture, and an expanded wrinkle texture for when the face is all about showing those expressions.

How are these textures applied, you might ask? Through a graphics engine that’s clever enough to use the tension map to apply these wrinkle textures to the 3D model. The result? An image rendered from the 3D model that looks so real, you’d think it’s a photograph.

However, this method isn’t just about creating one image. It can render many photos depicting the face with various wrinkles and expressions. These images can be used for training machine learning models, whether for a face image synthesizer, facial landmark recognition, or even predicting how you’ll age accurately.

The tech has a variety of applications, and it can be used to create realistic images of faces for video games, films, or any digital content.

You can read more about Microsoft’s wrinkle prediction tech here.

More about the topics: microsoft, Technology

User forum

0 messages