Srinivasa Narasimhan



 Assistant Professor
 The Robotics Institute
 School of Computer Science
 Carnegie Mellon University

Teaching

  • Computer Vision courses at Carnegie Mellon

  • 16-421: Vision Sensors [Spring 2009] [Spring 2008]
  • 16-823: Physics-based methods in Computer vision [Fall 2008]  [Fall 2007]  [Fall 2006]  [Fall 2005]
  • 15-385(685): Undergraduate Computer Vision [Spring 2007]  [Spring 2006]  (Carnegie Mellon Blackboard )


  • Seminars



    Group

  • Sanjeev J. Koppal (5th year, PhD, defending August 09)
  • Chenyu Wu (5th year, PhD, co-advisor, defending December 09)
  • Mohit Gupta (4th year, PhD)
  • Jean-Francois Lalonde (4th year, PhD, co-advisor)
  • Pete Barnum (3rd year, PhD, co-advisor)
  • Yuandong Tian (1st year, PhD)
  • Shuntaro Yamazaki (visiting researcher, AIST, Japan)
  • Jennifer Turken (support staff)


  • Research

    My research interests are in computer vision and computer graphics. Primarily, I am interested in developing physics-based models and algorithms for scene interpretation in vision and real-time rendering of physical effects in graphics. Selected projects are highlighted below. Please visit my official RI webpage for a list of my publications. The sponsors for these research projects include Office of Naval Research (ONR), Defense Advanced Projects Research Agency (DARPA), National Science Foundation (NSF), and Siemens Corporate Research, Princeton (SCR).



    Illumination and Imaging




    Temporal Dithering of Illumination

     

    (De)Focused Illumination and Global Light Transport




    Illumination from a single outdoor image

     

    What do the Sun and the Sky tell us about the Camera?




    Coplanar Shadowgram Imaging

     


    Reciprocal Imaging with Shadow Cameras (coming soon)

      • Description
      • Videos



    Display with Water-drops

     


    Imaging through water (coming soon)

      • Description
      • Videos

    Assorted Pixels: Multidimensional Imaging

     


    DLP Photography




    Light Transport for Computer Vision




    Structured Light in Scattering Media

     

    Optimal Placement of Source and Camera


    Frequency Space Analysis of Rain and Snow

     

    The Camera as a Weather Station

     
       

    Vision through Fog and Haze

     

    Weather and Illumination Database (WILD)

       


    Light Transport for Computer Graphics





    Legendre Fluids: Fast Modeling and Rendering of Participating Media

     



    Acquiring Scattering Properties of Participating Media by Dilution


    Single Scattering Model for Real-time Rendering

     

    Rendering Multiple Scattering Effects




    Appearance Analysis





    Bone Reconstruction for Orthopedic Endoscopy

     

    Surface Normal Clustering for Scene Analysis


    Novel Depth Cues from Uncalibrated Near-Field Lighting

     

    Photometric Invariants

       


    Grants

  • ONR N00014-08-1-0330, "Photometric Methods for Scene Understanding" [2008-2011]
  • US-Israel Binational Science Foundation, "Sensing Fusion for Underwater Scene Recovery"[2007-2010]
  • NSF CAREER IIS-0643628, "Making Computer Vision Successful in Scattering Media", [2007-2011]
  • ONR N00014-07-1-0522, "IEEE/ONR Workshop on Volumetric Scattering in Vision and Graphics", [2007-2007]
  • NSF CCF-0541307, "Collaborative Research: Fast and Accurate Volumetric Rendering of Scattering Phenomena in Computer Graphics", [2006-2008]
  • ONR DURIP N00014-06-1-0762, "Instrumentation in Support of Flexible Imaging Systems and Imaging in Dispersive Media", [2006-2007]
  • ONR N00014-05-1-0188, "A Physical Approach to Optical Underwater Imaging", [2005-2007]