Matting & Deblurring

Tuesday, 1 August
3:45 - 5:30 pm
Hall C
Session Chair: Paul Debevec, USC Centers for Creative Technologies

Flash Matting
Pick up a camera and a flash, extract hairy and furry things from photos. The Joint Bayesian Flash Matting is a powerful and practical natural image matting technique for everyone!

Jian Sun
Yin Li
Sing Bing Kang
Heung-Yeung Shum
Microsoft Research Asia


Natural Video Matting Using Camera Arrays
Using statistics derived from images from a camera array,this method computes alpha mattes for natural scenes with textured backgrounds. The algorithm is automatic and runs at near real-time rates.

Neel Joshi
University of California, San Diego

Wojciech Matusik
Shai Avidan
Mitsubishi Electric Research Laboratories (MERL)


Removing Camera Shake From a Single Photograph
Camera shake, in which an unsteady camera causes blurry photographs, is a chronic problem for photographers. This paper introduces an algorithm to remove these effects from seriously blurred images.

Rob Fergus
Barun Singh
Massachusetts Institute of Technology, Computer Science and Artificial Intelligence Laboratory

Aaron Hertzmann
Sam T. Roweis
University of Toronto

William T. Freeman
Massachusetts Institute of Technology, Computer Science and Artificial Intelligence Laboratory


Coded Exposure Photography: Motion Deblurring using Fluttered Shutter
This method flutters the camera shutter open and closed with a binary pseudo-random sequence during exposure. Fast-moving objects are blurred, but sequenced shuttering preserves all spatial frequencies, simplifying the deblurring process.

Ramesh Raskar
Amit Agrawal
Mitsubishi Electric Research Laboratories (MERL)

Jack Tumblin
Northwestern University