Abstract
In this paper the established link between Mathematical Morphology and DT-CNN is
used to derive a new template decomposition procedure. It is based on the decomposition
of morphological structuring elements. From a practical point of view, template
decomposition is an important issue since image processing operations can be non-local
while the cells in physical DT-CNN realizations often only communicate with cells in
their direct neighborhood. Here we apply techniques from morphological algebra to
derive a DT-CNN template decomposition strategy, which is far more efficient in terms
of the number of templates in the decomposition than the traditional decomposition
strategy. In a typical 9 x 9 example, the number of templates reduces from 125 to 14.
used to derive a new template decomposition procedure. It is based on the decomposition
of morphological structuring elements. From a practical point of view, template
decomposition is an important issue since image processing operations can be non-local
while the cells in physical DT-CNN realizations often only communicate with cells in
their direct neighborhood. Here we apply techniques from morphological algebra to
derive a DT-CNN template decomposition strategy, which is far more efficient in terms
of the number of templates in the decomposition than the traditional decomposition
strategy. In a typical 9 x 9 example, the number of templates reduces from 125 to 14.
Original language | English |
---|---|
Title of host publication | Proceedings BNAIC?02 |
Pages | 35-42 |
Publication status | Published - 2002 |
Event | Belgium/Netherlands Conference on Artificial Conference (BNAIC) - Leuven, Belgium Duration: 0001 Jan 2 → … |
Conference
Conference | Belgium/Netherlands Conference on Artificial Conference (BNAIC) |
---|---|
Country/Territory | Belgium |
City | Leuven |
Period | 0001/01/02 → … |
Subject classification (UKÄ)
- Electrical Engineering, Electronic Engineering, Information Engineering