Christian Häne



I am a postdoctoral scholar at the University of California, Berkeley under the guidance of Prof. Jitendra Malik. Before joining UC Berkeley I studied computer science at ETH Zürich, Switzerland. I received a BSc and MSc degree in Computer Science from ETH Zürich, Switzerland in 2010 and 2011, respectively. In 2016 I received a Dr. sc. degree from ETH Zürich under the supervision of Prof. Marc Pollefeys. My thesis was awarded with the ETH medal for outstanding doctoral theses. I am the recipient of an Early Postdoc.Mobility Fellowship from the Swiss National Science Foundation. My research interests include dense 3D reconstruction and the application of these methods to challenging scenarios, such as semantic 3D reconstruction and 3D reconstruction from single images. I am also interested in real-time applications such as self-driving cars.


  • September 2017: One paper accepted to 3DV and one paper accepted to NIPS
  • August 2017: Our tech report entitled "Learning a Multi-View Stereo Machine" is available
  • August 2017: Our submission to the IMAVIS special issue "Automotive Vision" got accepted
  • August 2017: Slides for our CVPR2017 tutorial are online
  • April 2017: New tech report about Hierarchical Surface Prediction is available
  • January 2017: I received the ETH Medal for outstanding doctoral theses
  • September 2016: Our paper "Dense Semantic 3D Reconstruction" was accepted for publication in TPAMI
  • September 2016: Two papers accepted at 3DV, one as oral
  • July 2016: Our paper about semantic 3D reconstruction of heads was accepted at ECCV
  • July 2016: I started as a postdoctoral scholar at the University of California, Berkeley
  • June 2016: We are giving a tutorial about semantic 3D modeling at CVPR. Tutorial material available here
  • February 2016: My plane sweeping stereo matching implementation is publicly available here


Semantic 3D Reconstruction

Dense Semantic 3D Reconstruction from multiple RGB input images

Single View 3D Reconstruction

3D Reconstruction from a single RGB input image

Reconstruction with Shape Priors

3D Reconstruction from RGB images using a multi-label head shape prior