Scheduled System Maintenance:
Some services will be unavailable Sunday, March 29th through Monday, March 30th. We apologize for the inconvenience.
By Topic

Joint reconstruction of compressed multi-view images

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

The purchase and pricing options are temporarily unavailable. Please try again later.
2 Author(s)
Xu Chen ; Electr. & Comput. Eng., Univ. of Illinois at Chicago, Chicago, IL ; Frossard, P.

This paper proposes a distributed representation algorithm for multi-view images that are jointly reconstructed at the decoder. Compressed versions of each image are first obtained independently with random projections. The multiple images are then jointly reconstructed by the decoder, under the assumption that the correlation between images can be represented by local geometric transformations. We build on the compressed sensing framework and formulate the joint reconstruction as a l2-l1 optimization problem. It tends to minimize the MSE distortion of the decoded images, under the constraint that these images have sparse and correlated representations over a structured dictionary of atoms. Simulation results with multi-view images demonstrate that our approach achieves better reconstruction results than independent decoding. Moreover, we show the advantage of structured dictionaries for capturing the geometrical correlation between multi-view images.

Published in:

Acoustics, Speech and Signal Processing, 2009. ICASSP 2009. IEEE International Conference on

Date of Conference:

19-24 April 2009