Segmentation and classification of leukocytes using neural networks: a generalization direction
In image digital processing, as in other fields, it is commonly difficult to simultaneously achieve a generalizing system and a specialistic system. The segmentation and classification of leukocytes is an application where this fact is evident. First an exclusively supervised approach to segmentatio...
| Autor principal: | |
|---|---|
| Outros Autores: | , |
| Formato: | bookPart |
| Idioma: | eng |
| Publicado em: |
2010
|
| Assuntos: | |
| Texto completo: | http://hdl.handle.net/10198/1445 |
| País: | Portugal |
| Oai: | oai:bibliotecadigital.ipb.pt:10198/1445 |
| Resumo: | In image digital processing, as in other fields, it is commonly difficult to simultaneously achieve a generalizing system and a specialistic system. The segmentation and classification of leukocytes is an application where this fact is evident. First an exclusively supervised approach to segmentation and classification of blood white cells images is shown. As this method produces some drawbacks related to the specialistic/generalized problems, another process formed by two neural networks is proposed. One is an unsupervised network and the other one is a supervised neural network. The goal is to achieve a better generalizing system while still doing well the role of a specialistic system. We will compare the performance of the two approaches. |
|---|