Utilize este identificador para referenciar este registo:
|Título:||A neuro-genetic approach to non-invasive temperature estimation in multilayered media under ultrasound therapy|
|Autor:||Teixeira, C. A.|
Ruano, M. Graça
Ruano, A. E.
Pereira, W. C. A.
|Citação:||Teixeira, C. A.; Ruano, M. G.; Ruano, A. E.; Pereira, W. C. A. A Neuro-Genetic Approach to Non-invasive Temperature Estimation in Multilayered Media under Ultrasound Therapy, Trabalho apresentado em 19th International Congress on Acoustics (ICA 2007), In Proceedings of the 19th International Congress on Acoustics (ICA 2007), Madrid, 2007.|
|Resumo:||Precise non-invasive temperature estimation is a primary aspect for a correct ultrasound therapy application. In multilayered media it is important to estimate the temperature inside each layer and especially at the interfaces, where higher temperature values can happen. In this work, it is proposed a method to estimate temperature from backscattered ultrasound signals (BSU) using genetically selected radial-basis functions neural networks (RBFNN). A three-layered gel-based phantom was build and submitted to 1-MHz ultrasound waves at several physiotherapeutic intensities. Low-intensity 5-MHz ultrasound pulses were also applied and backscattered and temperature signals were acquired. Temperatures were collected from thermocouples placed inside each layer and at their interfaces, along the axial line of the physiotherapeutic ultrasound transducer. After data collection, temporal echo-shifts due to temperature changes were computed from the BSU signals, and applied with the past temperature values as RBFNN inputs. After training and structure selection, the best RBFNN were validated in fresh data, and proper estimators which presented a maximum absolute error inferior to 0.5 ºC (standard threshold for hyperthermia/diathermia) were obtained. This precision was attained at both the spatial points and intensities. In addition, the best models present a small computational complexity, desired in future real-time applications.|
|Aparece nas colecções:||FCT2-Artigos (em revistas ou actas indexadas)|
Ficheiros deste registo:
|2007_ICA.pdf||388,41 kB||Adobe PDF||Ver/Abrir Acesso Restrito. Solicitar cópia ao autor!|
Todos os registos no repositório estão protegidos por leis de copyright, com todos os direitos reservados.