| Paper: | IMDSP-P12.10 | ||
| Session: | Image Coding | ||
| Time: | Friday, May 21, 13:00 - 15:00 | ||
| Presentation: | Poster | ||
| Topic: | Image and Multidimensional Signal Processing: Image and Video Coding | ||
| Title: | RIPPLE-FREE LOCAL BASES BY DESIGN | ||
| Authors: | John Lewis; University of Southern California | ||
| Zhenyao Mo; University of Southern California | |||
| Ulrich Neumann; University of Southern California | |||
| Abstract: | In some applications a local or ‘parts based’ representationis preferable to global basis functions such as thoseused in Fourier and principal component analysis. In applicationsthat require human understanding and editing of thedata, it is also desirable that the basis functions be in somesense as “simple” as possible. This means, for example,that the basis functions should not have Gabor-like ripplesif such ripples are not a prominent feature of the data to berepresented. This paper introduces a direct local basis construction.Specifically, we show that local bases result frommaximizing an appropriate redefinition of pairwise orthogonalitywhile maintaining the ability to represent the data.The resulting basis functions are competitive with (and forsome applications superior to) those obtained from existingalgorithms, and the construction does not require that thebasis coefficients be statistically non-Gaussian or independent,as would be the case with an independent componentanalysis approach. | ||
| Back | |||
Home -||-
Organizing Committee -||-
Technical Committee -||-
Technical Program -||-
Plenaries
Paper Submission -||-
Special Sessions -||-
ITT -||-
Paper Review -||-
Exhibits -||-
Tutorials
Information -||-
Registration -||-
Travel Insurance -||-
Housing -||-
Workshops