Shulou(Shulou.com)11/24 Report--

CTOnews.com, Nov. 16 (Xinhua)-- A new Microsoft patent, a new machine learning model patent that creates "more lifelike" avatars for users, was announced on the website of the U.S. Patent and Trademark Office on Tuesday, local time.

According to reports, through the new machine learning model, the avatar or photo can be adjusted for the details to make the photo look more natural. Microsoft will use a convolutional attention network to improve the accuracy of capturing facial expressions and adjust image details such as blood flow or blushing based on physiological signals such as heart rate.

Microsoft further described in the patent document that the "ultra-realistic" avatar can mimic not only blinking or head status, but also subtle changes such as blood flow, breathing or emotional response.

Foreign media mspoweruser analysis said that the patent may be landed in areas such as the creation of video game characters, of course, does not rule out being used in Microsoft's own Teams app "3D portrait". If the patent is eventually approved, it is expected to "change the rules of the game" and completely change people's perception of avatars.

According to previous reports from CTOnews.com, Microsoft announced the introduction of the Mesh platform for Microsoft Teams applications during the Build 2023 developer conference, and conference participants can use 3D avatars.

Microsoft said at the time that users of the Windows and macOS Teams apps could create 3D avatars of themselves so that they could be used in meetings without cameras or webcams.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.