Abstract
The problem of spectral factorization in twodimensions (2-D) is complicated by the lack of a fundamental theorem of algebra in 2-D. In this paper, the theoretical aspects of a new method of 2-D spectral factorization are presented. The Radon transform is used as a tool to reduce the 2-D problem to a set of independent 1-D problems, which can be solved by well developed 1-D techniques. The proposed method guarantees stability and correlation match, although in the Radon space, by virtue of being a set of 1-D problems. The method is useful for modeling and processing tomographic data.
Original language | English |
---|---|
Pages | 15-18 |
Number of pages | 4 |
DOIs | |
Publication status | Published - 01-01-1994 |
Event | 7th IEEE SP Workshop on Statistical Signal and Array Processing, SSAP 1994 - Quebec, Canada Duration: 26-06-1994 → 29-06-1994 |
Conference
Conference | 7th IEEE SP Workshop on Statistical Signal and Array Processing, SSAP 1994 |
---|---|
Country/Territory | Canada |
City | Quebec |
Period | 26-06-94 → 29-06-94 |
All Science Journal Classification (ASJC) codes
- Statistics, Probability and Uncertainty
- Signal Processing