Manmohan Chandraker

Manmohan Chandraker

I lead the Media Analytics group at NEC Labs America. We work on computer vision, in particular large-scale image recognition and 3D scene understanding. My personal research interests are 3D reconstruction and graphics-based vision, with applications in autonomous driving, robotics and HCI.

My PhD was from UC San Diego with Prof. David Kriegman, before a postdoc at UC Berkeley with Prof. Ravi Ramamoorthi.

E-mail: manu [AT] nec - labs . com

Bio

CV

Papers

Honors

NEWS

2016: Three papers accepted for CVPR 2016 on weakly-supervised reconstruction, shape from light fields and occlusion modeling.

2016: I will serve as area chair for ICVGIP 2016.

2016: I will serve as tutorials and workshop chair for 3DV 2016.

2016: Paper on scalable traffic scene recognition accepted at WACV 2016.

2015: Paper on BRDF-invariant shape recovery accepted for PAMI special issue of best papers from CVPR 2014.

2015: Paper on low scale-drift real-time monocular SFM accepted for IEEE PAMI.

2015: I will serve as area chair at CVPR 2016.

2015: Paper on 3D localization accepted as oral at CVPR 2015.

2015: I will serve as area chair for ICVGIP 2015.

2014: Work on BRDF-invariance with camera motion wins Best Paper Award at CVPR 2014.

2014: Paper on joint shape and material recovery accepted at ECCV 2014.

2014: Two papers accepted at CVPR 2014 on BRDF-invariance and real-time monocular SFM.

SELECTED CURRENT RESEARCH

3D Visual Sensing for Automated Driver Assistance: Our team is developing a real-time traffic scene understanding system that uses a single camera to localize traffic participants in 3D, for applications like collision avoidance and dangerous event prediction.

3D Object Localization
CVPR 2015 (oral)

Real-time Monocular SFM
Architecture: ICRA 2013 Scale-drift: CVPR 2014

Real-time Line SFM
ICCV 2009

Dense 3D Reconstruction for Challenging Conditions: My research aims to overcome the limitations of traditional dense reconstruction methods by harnessing semantic information, or with a deeper understanding of complex material behavior.

Stratified Autocalibration, Marr Prize 2007

Multiview stereo for complex materials
CVPR 2014 (oral)
[Best Paper Award]

Optical flow for unknown material type
CVPR 2013 (oral)

Dense reconstruction with semantic priors
CVPR 2013 (oral)

Material recovery using a single image
ICCV 2011 (oral)

Photometric stereo for unknown material type
CVPR 2011 (oral)

SELECTED PAST RESEARCH

Global Optimization for 3D Reconstruction: My PhD thesis is on globally optimal algorithms for structure and motion estimation. [PDF]
[Winner, CSE Dissertation Award for Best Doctoral Thesis, UCSD, 2009.] [Nominee, ACM Dissertation Award, 2009.]

Stratified Autocalibration
ICCV 2007 (oral)
[Marr Prize Hon. Mention for Best Paper]

Bilinear Programming
CVPR 2008 (oral)

Direct Autocalibration
CVPR 2007

Multiple-View Geometry
ECCV 2006 (oral)

Global Light Transport: I have developed theory and algorithms for photometric 3D reconstruction that can handle global light transport effects, such as inter-reflections and shadows. I have also worked on methods to invert such global effects in projector-camera systems.

Inter-reflections resolve GBR
CVPR 2005 (oral)

Photometric stereo with shadows
CVPR 2007

Inverse light transport
ECCV 2010

PUBLICATIONS

A. Kanazawa, D, Jacobs and M.K. Chandraker
WarpNet: Weakly Supervised Matching for Single-View Reconstruction
CVPR 2016, Las Vegas, NV.

T.-C. Wang, M.K. Chandraker, A. Efros and R. Ramamoorthi [oral]
SVBRDF-Invariant Shape and Reflectance Recovery from Light Fields
CVPR 2016, Las Vegas, NV.

V. Dhiman, Q.-H. Tran, J. Corso and M.K. Chandraker
A Continuous Occlusion Model for Road Scene Understanding
CVPR 2016, Las Vegas, NV.

C.-Y. Chen, W. Choi and M.K. Chandraker
Atomic Scenes for Scalable Traffic Scene Recognition
WACV 2016, Lake Placid, NY.

M.K. Chandraker [Spl. Issue, Best of CVPR 2014]
The Information Available to a Moving Observer on Shape with Unknown, Isotropic BRDFs
IEEE PAMI 2015.

S. Song, M.K. Chandraker and C. Guest
High Accuracy Monocular SFM and Scale Correction for Autonomous Driving
IEEE PAMI 2015.

S. Song and M.K. Chandraker [oral]
Joint SFM and Detection Cues in 3D Object Localization for Autonomous Driving
CVPR 2015, Boston, Massachusetts.

M.K. Chandraker
On Joint Shape and Material Recovery from Motion Cues [PDF]
ECCV 2014, Zurich, Switzerland.

M.K. Chandraker [oral] [Best Paper Award]
What Camera Motion Reveals About Shape with Unknown BRDF [PDF] [Tech Report]
CVPR 2014, Columbus, Ohio.

S. Song and M.K. Chandraker
Robust Scale Estimation in Real-Time Monocular SFM for Autonomous Driving [PDF]
CVPR 2014, Columbus, Ohio.

M.K. Chandraker, D. Reddy, Y. Wang and R. Ramamoorthi [oral]
What Motion Reveals About Shape with Unknown BRDF and Lighting [PDF]
CVPR 2013, Portland, Oregon.

Y. Bao, M.K. Chandraker, Y. Lin and S. Savarese [oral]
Dense Object Reconstruction with Semantic Priors [PDF]
CVPR 2013, Portland, Oregon.

S. Song and M.K. Chandraker
Parallel, Real-Time Monocular Visual Odometry [PDF]
ICRA 2013, Karlsruhe, Germany.

M.K. Chandraker, J. Bai and R. Ramamoorthi [Spl. Issue, Best of CVPR 2011]
On Differential Photometric Reconstruction for Unknown, Isotropic BRDFs [PDF]
IEEE PAMI 35(12):2941-2955, December 2013.

M.K. Chandraker and R. Ramamoorthi [oral]
What An Image Reveals About Material Reflectance [PDF]
ICCV 2011, Barcelona, Spain.

M.K. Chandraker, J. Bai and R. Ramamoorthi [oral]
A Theory of Photometric Reconstruction for Unknown Isotropic Reflectances [PDF]
CVPR 2011, Colorado Springs, pp. 2505-2512.

M.K. Chandraker, J. Bai, T.-T. Ng and R. Ramamoorthi
On the Duality of Forward and Inverse Light Transport [PDF]
IEEE PAMI 33(10):2122-2128, October 2011.

M.K. Chandraker, S. Agarwal, D.J. Kriegman and S. Belongie
Globally Optimal Algorithms for Stratified Autocalibration [PDF]
IJCV 90(2):236-254, November 2010.

J. Bai, M.K. Chandraker, T.-T. Ng and R. Ramamoorthi
A Dual Theory of Inverse and Forward Light Transport [PDF] [Tech Report] [Web]
ECCV 2010, Heraklion, Greece.

M.K. Chandraker, J. Lim and D.J. Kriegman
Moving in Stereo: Efficient Structure and Motion Using Lines [PDF]
ICCV 2009, Kyoto, Japan.

M.K. Chandraker and D.J. Kriegman [oral]
Globally Optimal Bilinear Programming for Computer Vision Applications [PDF] [Code]
CVPR 2008, Anchorage, Alaska.

F. Kahl, S. Agarwal, M.K. Chandraker, D.J. Kriegman and S. Belongie
Practical Global Optimization for Multiview Geometry [PDF] [Code]
IJCV 79(3):271-284, September 2008.

M.K. Chandraker, S. Agarwal, D.J. Kriegman and S. Belongie [oral] [Marr Prize Honorable Mention]
Globally Optimal Affine and Metric Upgrades in Stratified Autocalibration [PDF] [Web]
ICCV 2007, Rio de Janeiro, Brazil.

A. Agarwal, S. Izadi, M.K. Chandraker and A. Blake
High Precision Multi-touch Sensing on Surfaces using Overhead Cameras [PDF]
IEEE Tabletop 2007, Rhode Island.

M.K. Chandraker, S. Agarwal and D.J. Kriegman
ShadowCuts: Photometric Stereo with Shadows [PDF] [Data+Code]
CVPR 2007, Minneapolis, Minnesota.

M.K. Chandraker, S. Agarwal, F. Kahl, D. Nistér and D.J. Kriegman
Autocalibration via Rank-Constrained Estimation of the Absolute Quadric [PDF] [VRML]
CVPR 2007, Minneapolis, Minnesota.

S. Agarwal, M.K. Chandraker, F. Kahl, D.J. Kriegman and S. Belongie [oral]
Practical Global Optimization for Multiview Geometry [PDF] [Code]
ECCV 2006, Graz, Austria, pp. 592-605, vol. 1.

M.K. Chandraker, F. Kahl and D.J. Kriegman [oral]
Reflections on the Generalized Bas-Relief Ambiguity [PDF]
CVPR 2005, San Diego, CA, pp. 788-795, vol. 1.

M.K. Chandraker, C. Stock and A. Pinz
Real-Time Camera Pose in a Room
ICVS 2003, Graz Austria, pp. 98-110.

C. Stock, U. Mühlmann, M.K. Chandraker and A. Pinz
Subpixel Corner Detection for Tracking Applications using CMOS Camera Technology [PDF]
AAPR 2002, Graz, Austria.