Profile cover photo
Profile photo
Michael Black
474 followers -
Computer vision scientist and director at the Max Planck Institute for Intelligent Systems
Computer vision scientist and director at the Max Planck Institute for Intelligent Systems

474 followers
About
Michael's posts

Post has attachment
Body Labs has released "mosh" full body "filters" -- the app takes a single image, estimates the 3D body shape and pose, and lets users apply all kinds of fun filters that know about the body. It is available the Apple app store. This is just a first step towards computers that really "see" us. Give it a try and let me know what you think! https://www.bodylabs.com/mosh/

Post has attachment

Post has attachment

Post has attachment

Post has attachment
Create accurate 3D bodies without a scanner using verbal descriptions. Our Body Talk system is being presented this morning at SIGGRAPH -- first session, Ballroom D.

Post has attachment

Post has attachment
Here is a preview of our CVPR'16 paper, which uses semantic image segmentation to improve optical flow estimation, with top results among monocular methods on KITTI 2015.  https://ps.is.tuebingen.mpg.de/publications/sevilla-cvpr-2016

Post has attachment
That's leadership.  It really makes the point that programming is power today.  It is part of a general education.  Nicely done.

Post has attachment
Our new SIGGRAPH Asia paper defines a 3D body model that is more accurate than our previous models, is compatible with existing graphics systems, and is available for research purposes. The website and model will be available in the next few days.

Post has attachment
At ICCV'11 we introduced the first method to capture human body shape and pose using a Kinect. For ICCV'15 we have significantly improved on this by capturing accurate and highly detailed shape using RGB-D sequences.w
Wait while more posts are being loaded