Hierarchical description and analysis of CNN algorithms
When designing algorithms, several implementations are often feasible, performing differently on the hardware. We introduce a formalism to hierarchically describe CNN-based algorithms, aimed at an execution on a CNN Universal Machine. Apart from the mere description, the formalism facilitates perfor...
Saved in:
| Published in | Proceedings (International Workshop on Cellular Neural Networks and Their Applications) pp. 1 - 2 |
|---|---|
| Main Authors | , , |
| Format | Conference Proceeding |
| Language | English |
| Published |
IEEE
01.07.2014
|
| Subjects | |
| Online Access | Get full text |
| ISSN | 2165-0144 |
| DOI | 10.1109/CNNA.2014.6888629 |
Cover
| Summary: | When designing algorithms, several implementations are often feasible, performing differently on the hardware. We introduce a formalism to hierarchically describe CNN-based algorithms, aimed at an execution on a CNN Universal Machine. Apart from the mere description, the formalism facilitates performance estimation and programming of the algorithms. |
|---|---|
| ISSN: | 2165-0144 |
| DOI: | 10.1109/CNNA.2014.6888629 |