My research interests lie at the intersection of computer graphics, machine learning, computer vision, and HCI. Recently, I have started to focus on bringing the 2D and 3D worlds together by bridging the gap between different media modalities: 3D shapes, video, images and depth scans. Visit my publications page for more details.
My PhD work was on developing algorithms to infer high-level semantic and functional attributes of shapes to build intelligent design tools. I focused on developing frameworks where the end user is abstracted from low-level interaction with the content, but is enabled with simple and intuitive 'natural language like' controllers to manipulate the geometry and appearance of it.
Internship Opportunities @ Adobe Research
If you are a PhD student interested in doing a research internship at Adobe and/or collaborating on a project with me, send an e-mail. I do get a significant amount of such requests, hence it will be very helpful if you include a CV, your affiliation, your advisor's name, and a summary of of your current research interests.
yumer [at] adobe (dot) com
meyumer [at] gmail (dot) com
Learning Volumetric 3D Object Reconstruction from Single-View with Projective Transformations paper is accepted for publication at NIPS 2016.
Learning Semantic Deformation Flows with 3D Convolutional Networks paper is accepted for publication at ECCV 2016.
Spectral Style Transfer for Human Motion between Independent Actions paper is conditionally accepted for publication at SIGGRAPH 2016.