Image Analysis & Enhancement

Monday, 6 August
8:30 - 10:15 am
Room 6AB & CF
Session Chair/Discussant: Richard Szeliski, Microsoft Research

Image Deblurring With Blurred/Noisy Image Pairs

A novel method to deblur an image, with the help of another noisy image of the same scene. The paper also proposes new methods of de-ringing the deconvoluted image.

Lu Yuan
The Hong Kong University of Science and Technology

Jian Sun
Microsoft Research Asia

Long Quan
The Hong Kong University of Science and Technology

Heung-Yeung Shum
Microsoft Research Asia

Solid Texture Synthesis From 2D Exemplars

A novel method for synthesizing solid textures from 2D exemplars. In addition to producing compelling texture-mapped surfaces, this method models the material in the interior of solid objects.

Johannes Kopf
Universität Konstanz

Chi-Wing Fu
The Hong Kong University of Science and Technology

Daniel Cohen-Or
Tel Aviv University

Oliver Deussen
Universität Konstanz

Dani Lischinski
The Hebrew University

Tien-Tsin Wong
The Chinese University of Hong Kong

Photo Clip Art

An interactive system for inserting new objects into existing photographs using a vast image-based object library.

Jean-François Lalonde
Derek Hoiem
Alexei A. Efros
Carnegie Mellon University

Carsten Rother
John Winn
Antonio Criminisi
Microsoft Research Cambridge

Scene Completion Using Millions of Photographs

A method for filling holes in images using semantically and structurally similar scenes from a large internet photo library.

James Hays
Alexei A. Efros
Carnegie Mellon University