Abstract
This paper presents a novel approach to accelerate the 2D Mallat decomposition algorithm. In particular the proposed approach performs the 2D Mallat decomposition algorithm by way of computing the 2D cyclical convolution, and further completes the 2D cyclical convolution with the fast number theoretical transform (FNTT). Therefore, the 2D Mallat decomposition algorithm can be speeded up. The results obtained from the theoretical analysis and the experiment show that the 2D Mallat decomposition algorithm is significantly improved.
Original language | English |
---|---|
Title of host publication | Proceedings of the Fifth International Conference on Document Analysis and Recognition. ICDAR '99 (Cat. No.PR00318) |
Editors | Bob Werner |
Publisher | IEEE Canada |
Pages | 87-90 |
Number of pages | 4 |
ISBN (Print) | 0769503187 |
DOIs | |
Publication status | Published - 20 Sept 1999 |
Externally published | Yes |
Event | Proceedings of the Fifth International Conference on Document Analysis and Recognition. ICDAR '99 (Cat. No.PR00318) - Bangalore, India Duration: 22 Sept 1999 → 22 Sept 1999 |
Conference
Conference | Proceedings of the Fifth International Conference on Document Analysis and Recognition. ICDAR '99 (Cat. No.PR00318) |
---|---|
Period | 22/09/99 → 22/09/99 |
User-Defined Keywords
- Acceleration
- Convolution
- Signal processing algorithms
- Low pass filters
- Sun
- Computer science
- Mathematics
- Read only memory
- Digital signal processing
- Image processing