Automatic registration of digital images is an important support in the medical field for physicians and surgeons. In fact, comparison of anatomical scan is a fundamental procedure for disease prediction, lesions quantification or for evaluating the results of a therapy. A new proposed approach implements three-dimensional neural networks to match, and hence to register, volumetric data sets of the brain in order to evaluate the differences between two volumes. The high computational complexity of this approach has been improved by implementing a more efficient method to train the networks.
Computational complexity analysis of a 3D neural network approach to volume matching
Salvetti O;
2002
Abstract
Automatic registration of digital images is an important support in the medical field for physicians and surgeons. In fact, comparison of anatomical scan is a fundamental procedure for disease prediction, lesions quantification or for evaluating the results of a therapy. A new proposed approach implements three-dimensional neural networks to match, and hence to register, volumetric data sets of the brain in order to evaluate the differences between two volumes. The high computational complexity of this approach has been improved by implementing a more efficient method to train the networks.File in questo prodotto:
File | Dimensione | Formato | |
---|---|---|---|
prod_43653-doc_122568.pdf
solo utenti autorizzati
Descrizione: Computational complexity analysis of a 3D neural network approach to volume matching
Tipologia:
Versione Editoriale (PDF)
Dimensione
865.28 kB
Formato
Adobe PDF
|
865.28 kB | Adobe PDF | Visualizza/Apri Richiedi una copia |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.