RI | Research | Projects | Scene Flow

Text only version of this site

Scene Flow This project is no longer active. Head: Simon Baker Contact: Simon Baker Mailing address: Carnegie Mellon University Robotics Institute 5000 Forbes Avenue Pittsburgh, PA 15213 Associated center: VASC Associated labs/groups: Vision for Virtual Environments and Virtualized RealityTM

Jump to: Project Description | Personnel | Publications

Project Description

Scene flow is the three-dimensional motion field of points in the world, just as optical flow is the two-dimensional motion field of points in an image. Any optical flow is simply the projection of the scene flow onto the image plane of a camera. We have developed a framework for the computation of dense, non-rigid scene flow from optical flow. We have also developed an algorithm for computing 3D scene flow and 3D scene shape simultaneously. This second algorithm does not require the prior computation of optical flow. See the movie below for an example of a reconstructed 3D voxel model and the corresponding 3D scene flow. As the movie plays, the voxels in the 3D are "flowed" along the scene flow.

Image shows video sequence.

See our Spatio-Temporal View Interpolation algorithm for a better way of rendering such scenes across both space and time.

Past members

	Name	Title	Email Address
	Simon Baker	Adjunct Faculty (Adjunct)
	Robert Collins	Adjunct Faculty (Adjunct)	rcollins@andrew.cmu.edu
	Takeo Kanade	U.A. and Helen Whitaker University Prof.	tk@cs.cmu.edu
	Arvind Lakshmikumar	Visiting Researcher
	Peter Rander	Senior Commercialization Specialist	rander@cs.cmu.edu
	Steven Seitz	Adjunct Faculty (Adjunct)
	Sundar Vedula	PhD Student, RI

Publications

Note: This list may not be comprehensive. It contains only those publications in the RI publications database. Entries are listed in reverse chronological order.

• Image-Based Spatio-Temporal Modeling and View Interpolation of Dynamic Events
  S. Vedula, S. Baker, and T. Kanade
  ACM Transactions on Graphics, Vol. 24, No. 2, April, 2005, pp. 240 - 261. [Abstract]
  Download: pdf [2053 KB] copyrighted

• Three-Dimensional Scene Flow
  S. Vedula, S. Baker, P. Rander, R. Collins, and T. Kanade
  IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 27, No. 3, March, 2005, pp. 475 - 480. [Abstract]
  Download: pdf [576 KB] copyrighted

• Spatio-Temporal View Interpolation
  S. Vedula, S. Baker, and T. Kanade
  Proceedings of the 13th ACM Eurographics Workshop on Rendering, June, 2002. [Abstract]
  Download: pdf [2394 KB], ps.gz [19353 KB] copyrighted

• Image Based Spatio-Temporal Modeling and View Interpolation of Dynamic Events
  S. Vedula
  doctoral dissertation, tech. report CMU-RI-TR-01-37, Robotics Institute, Carnegie Mellon University, September, 2001. [Abstract]
  Download: pdf [5514 KB] copyrighted

• Shape and Motion Carving in 6D
  S. Vedula, S. Baker, S. Seitz, and T. Kanade
  Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR '00), June, 2000. [Abstract]
  Download: pdf [484 KB], ps.gz [517 KB] copyrighted

• Three-Dimensional Scene Flow
  S. Vedula, S. Baker, P. Rander, R. Collins, and T. Kanade
  Proceedings of the 7th International Conference on Computer Vision, Vol. 2, September, 1999, pp. 722 - 729. [Abstract]
  Download: pdf [412 KB], ps.gz [628 KB] copyrighted

• Multi-view Spatial and Temporal Interpolation for Dynamic Event Visualization
  S. Vedula
  tech. report CMU-RI-TR-99-15, Robotics Institute, Carnegie Mellon University, June, 1999. [Abstract]
  Download: pdf [975 KB], ps.gz [1768 KB] copyrighted