Utilize este identificador para referenciar este registo: http://hdl.handle.net/10400.1/2223
Título: Regressive and non-regressive RBFNN estimators in non-invasive temperature estimation
Autor: Teixeira, C. A.
Ruano, M. Graça
Ruano, A. E.
Pereira, W. C. A.
Palavras-chave: Non-invasive Temperature Estimation
Neural Networks
Multi-objective Optimization
Therapeutic Ultrasound
Data: 2008
Citação: Teixeira, C. A.; Pereira, W. C. A.; Ruano, M. G.; Ruano, A. E. Regressive and non-regressive RBFNN estimators in non-invasive temperature estimation, Trabalho apresentado em 21º Congresso Brasileiro de Engenharia Biomédica, In Actas do 21º Congresso Brasileiro de Engenharia Biomédica, Baía, 2008.
Resumo: In this paper two modelling strategies were applied for non-invasive temperature estimation in a gel-based phantom subjected to physiotherapeutic ultrasound. The two strategies differ in the consideration of regressive or nonregressive radial basis functions neural networks (RBFNN) structures. The gel-based phantom was heated using four different ultrasound intensities. Temperature was monitored at five points inside the phantom, where temperature is to be estimated. The best regressive model reaches a mean maximum absolute error of 0.4 ºC, against 0.8 ºC presented by the best non-regressive model. The regressive strategy presented better error performance with a smaller computational complexity. Thus, being the appropriate one for real-time temperature monitoring in hyperthermia/diathermia procedures, where resolutions below 0.5 ºC are required.
Peer review: yes
URI: http://hdl.handle.net/10400.1/2223
ISBN: 978-85-60064-13-7
Aparece nas colecções:FCT2-Artigos (em revistas ou actas indexadas)

Ficheiros deste registo:
Ficheiro Descrição TamanhoFormato 
CBEB2008_0280_Cesar_RBFNN_Temperature.pdf503,58 kBAdobe PDFVer/Abrir    Acesso Restrito. Solicitar cópia ao autor!


FacebookTwitterDeliciousLinkedInDiggGoogle BookmarksMySpace
Formato BibTex MendeleyEndnote 

Todos os registos no repositório estão protegidos por leis de copyright, com todos os direitos reservados.