#科技头条#【一张自拍就能创建三维头像 AI技术即将开启3D社交时代?】

#科技头条#【一张自拍就能创建三维头像  AI技术即将开启3D社交时代?】加州初创公司推出智能 3D 模型软件 Loom.ai ,利用机器学习和计算机视觉技术,根据用户 2D 平面照片,生成 3D 头像,而且这些头像都有着随时可以变化的情绪。据悉,未来该技术将会应用于社交软件、虚拟现实游戏及电子商务中,其中包括Facebook。

585231a04494b

Transform your selfie into a 3D avatar: Startup uses computer vision and machine learning to create life-like human faces

Forget enhancing your selfies with filters, now you can transform them into expressive 3D avatars.

Loom.ai generates detailed 3D avatars from basic 2D photos using machine learning and computer vision.

These creations can be shared in text messages, virtual reality games, on social media and e-commerce sites.

'We achieve similar visual fidelity as a scan, yet fully animatable in 3D,' Loom.ai shares on its website.

'We also factor out the lighting from the photograph, which means that your 3D avatar can be immersed into new environments with different lighting.'

'We are talking about bringing your avatar to life with a AAA-grade facial musculature rig that has been adapted to fit your face.

'This could lead to synthesizing new animation and dynamic interactions in 3D virtual worlds.'

Loom.ai, although still in beta, is the brain child of visual effects artists Mahesh Ramasubramanian and Kiran Bhat.

Ramasubramanian previously worked as a visual effects supervisor at Dream Works Animation for movies like Shrek and Madagascar 3.

And Bhat worked on facial animation at LucasFilm for flicks The Avengers and Pirates of the Caribbean.

'We believe that you need to bring the user into the experience, which means figuring out a way to digitize the everyday user just like we used to digitize the movie stars,' Bhat told Fortune.

The 3D avatar faces are complete with emotions that can be seen in their eyes mouth and the movement of their face.

They can appear sad, happy, interested or even angry - however, there is only a small set of emotions available at the moment.

Users just have to upload a font facing image of themselves or someone else to the Loom.ai website, once it is out of beta – and the software does the rest.

Machine learning captures the details of the human's face and artificial intelligence fills in any areas that may be missing from the 2D image.

And the more selfies this software is fed, the better it will become at guessing other facial expression in the future.

It takes about a minute to build the avatar model and another minute to build the personalized facial rig with animation.

The system uses Sketchfab to view the results and it takes about 5 minutes to upload to the online site where it is shared for users to see.

The firm is also working on technology that will let users change the hairstyle and eye color of their avatar, and they will be able to add accessories like glasses.

Users interested in seeing what they look like in 3D can sign a waiting list and Loom.ai will notify you once they've moved onto their alpha stage of development.

The California startup plans to license the technology to third-party platforms that will incorporate it into new or existing applications.

And although Loom.ai would not tell Fortune what third-part platforms are testing its API, they did say they are looking at those with an extensive user base, like Facebook.

'Facebook has a ton of users, and we would love it if they would use our technology to turn a 2D photograph into a 3D avatar in the future,' Ramasubramanian says.

In October, Facebook unveiled its first virtual reality version of its software.

The new software allows people to join the same virtual space, and do everything from watch a video to be transported to the surface of Mars.

http://www.dailymail.co.uk/sciencetech/article-4035046/Transform-selfie-3D-avatar-Startup-uses-computer-vision-machine-learning-create-life-like-human-faces.html


Comments are closed.



无觅相关文章插件