ICIP 2006, Atlanta, GA
 

Slide Show

Atlanta Conv. & Vis. Bureau

 

Technical Program

Paper Detail

Paper:WA-L1.6
Session:Error Resiliance and Concealment
Time:Wednesday, October 11, 11:40 - 12:00
Presentation: Lecture
Title: A JOINT MOTION-IMAGE INPAINTING METHOD FOR ERROR CONCEALMENT IN VIDEO CODING
Authors: Liyong Chen; University of Hong Kong 
 Shing-Chow Chan; University of Hong Kong 
 Heung-Yeung Shum; Microsoft Research Asia 
Abstract: In this paper, we propose a new method for spatial-temporal error concealment in video coding using joint motion-image inpainting. The proposed method combines motion inpainting and adaptive Markov random field (MRF) based diffusion as robust motion inpainting. Image inpainting is employed to refine the result. With robust motion inpainting, effective compromise between temporal and spatial method is achieved for each point in the missing marcroblock (MB), which provides satisfactory visual quality of the restored frame with complex motion in video.