Björn Stenger

Lead Research Scientist
Rakuten Institute of Technology (RIT)

E-mail         Scholar         Publications

About me

I am a researcher at the Rakuten Institute of Technology, where I work on computer vision and machine learning.

Prior to Rakuten I was at Toshiba Research (archived page). I completed my PhD in computer vision at the University of Cambridge in Roberto Cipolla's group, co-supervised by Philip Torr. My Diplom in Computer Science is from the University of Bonn. During my studies, I spent time at the University of Victoria, Hewlett-Packard, and Siemens Corporate Research.

News

2018/06 - Attending CVPR 2018, bid for CVPR 2022 in New Orleans

2018/05 - Welcome Miao Jiang as a summer intern!

2018/03 - Kelvin Cheng presents Mixed Reality Experience at the FC Barcelona Museum.

2018/02 - Code for image super-resolution from ICIP 2017 ByNet paper available here.

Recent work

Depth-Based 3D Hand Pose Estimation: From Current Achievements to Future Goals
S. Yuan, G. Garcia-Hernando, B. Stenger, G. Moon, J. Y. Chang, K. M. Lee, P. Molchanov, J. Kautz, S. Honari, L. Ge, J. Yuan, X. Chen, G. Wang, F. Yang, K. Akiyama, Y. Wu, Q. Wan, M. Madadi, S. Escalera, S. Li, D. Lee, I. Oikonomidis, A. Argyros, and T.-K. Kim,
CVPR, June 2018.

Analysis of the top performing methods in the Hands in the Million Challenge.

[arxiv] [bibtex]
@InProceedings{YuanCVPR2018,
  author =   {Yuan, S. and Garcia-Hernando, G. and Stenger, B. and  Moon, G. 
              and  Chang, J. Y. and Lee, K. M. and  Molchanov, P. and 
              Kautz, J. and Honari, S. and  Ge, L. and  Yuan, J. and  
              Chen, X. and Wang, G. and Yang, F. and Akiyama, K. and Wu, Y. 
              and Wan, Q. and Madadi, M. and Escalera, S. and Li, S. and 
              Lee, D. and Oikonomidis, I. and Argyros, A. and Kim, T.-K.},
  title =    {3D Hand Pose Estimation: From Current Achievements 
              to Future Goals},
  booktitle = {CVPR},
  year =      {2018},
  month =     {June}
}
    
ByNet-SR: Image Super Resolution with a Bypass Connection Network
J. Xu, Y. Chae, B. Stenger,
ICIP, September 2017.

CNN-based image super-resolution with high quality results and fast run-time.

[bibtex][code]
@InProceedings{XuICIP2017,
  author = {Xu, J. and Chae, Y. and Stenger, B.},
  title  = {{ByNet-SR}: Image Super Resolution with a 
            Bypass Connection Network},
  year   = {2017},
  month = {September},
  booktitle = {ICIP}
}
    
BigHand2.2M Benchmark: Hand Pose Data Set and State of the Art Analysis
S. Yuan, Q. Ye, B. Stenger, S. Jain, T.-K. Kim,
CVPR, July 2017.

We captured a new standard dataset for 3D hand pose estimation. Depth maps are accurately annotated with 3D joint locations using a magnetic tracking system. We show that training a CNN on this data achieves accurate results. The data was used in the Hands in the Million Challenge.

[arxiv] [bibtex]

@InProceedings{YuanCVPR2017,
  author =   {Yuan, S. and Ye, Q. and Stenger, B. 
              and Jain, S. and Kim, T.-K."},
  title =    {BigHand2.2M Benchmark: Hand Pose Data Set 
              and State of the Art Analysis},
  booktitle = {CVPR},
  year =      {2017},
  month =     {July}
}
    
Parsing Floor Plan Images
S. Dodge, J. Xu, B. Stenger,
MVA, May 2017.

Wall segmentation using fully convolutional networks (FCN) and applications in furniture fitting and 3D modeling

[R-FP dataset] [video] [bibtex]
@InProceedings{DodgeMVA2017,
   author  = {Dodge, S. and Xu, J. and Stenger, B.},
   title   = {Parsing Floor Plan Images},
   year    = {2017},
   month   = {May},
   booktitle = {MVA}
}
    
Pano2CAD: Room Layout From A Single Panorama Image
J. Xu, B. Stenger, T. Kerola, T. Tung,
WACV, March 2017.

Estimating 3D room geometry from a single panorama image using surface normal estimation, 2D object detection and 3D object pose estimation

[arxiv] [bibtex]
@InProceedings{XuWACV2017,
   author  = {Xu, J. and Stenger, B. and Kerola T. and Tung, T.},
   title   = {Pano2CAD: Room Layout From A Single Panorama Image},
   year    = {2017},
   month   = {March},
   booktitle = {WACV}
}
    
design credits