 |
Technical Program
Paper Detail
| Paper: | MP-L2.8 |
| Session: | Distributed Image and Video Coding - II |
| Time: | Monday, October 9, 17:00 - 17:20 |
| Presentation: |
Lecture
|
| Title: |
A MODEL-BASED APPROACH TO CORRELATION ESTIMATION IN WAVELET-BASED DISTRIBUTED SOURCE CODING WITH APPLICATION TO HYPERSPECTRAL IMAGERY |
| Authors: |
Ngai-Man Cheung; University of Southern California | | | | | Antonio Ortega; University of Southern California | | |
| Abstract: |
Coding efficiency of many distributed source coding (DSC) applications depends on the accuracy of correlation estimation, which often has to be performed under rate and complexity constraints. In this work we propose to tackle correlation estimation in wavelet-based DSC by extending our previously proposed model-based estimation. We need to address two issues for the extension: (i) bit-planes are typically generated by sophisticated algorithms in wavelet-based DSC; (ii) certain wavelet subbands may not have enough coefficients for reliable estimation. We propose solutions to these problems and, using a hyperspectral image system as example, we demonstrate model-based estimation can lead to efficient implementation with lower computational and data exchange requirements, and improved parallelism, while incurring only small degradation in coding efficiency. |
|
