Repository logo
 
Loading...
Profile Picture

Search Results

Now showing 1 - 3 of 3
  • Single black-box models for two-point non-invasive temperature prediction
    Publication . Teixeira, C. A.; Ruano, M. Graça; Ruano, Antonio; Pereira, W. C. A.; Negreira, C.
    In this paper the performance of a genetically selected radial basis functions neural network is evaluated for non-invasive two-point temperature estimation in a homogeneous medium, irradiated by therapeutic ultrasound at physiotherapeutic levels. In this work a single neural network was assigned to estimate the temperature profile at the two considered points, and more consistent results were obtained than when considering one model for each point. This result was possible by increasing the model complexity. The best model predicts the temperature from two unseen data sequences during approximately 2 hours, with a maximum absolute error less than 0.5 oC, as desired for a therapeutic temperature estimator.
  • A data-driven methodology for non-invasive tempearture estimation in multilayered media
    Publication . Teixeira, C. A.; Pereira, W. C. A.; Ruano, M. Graça; Ruano, Antonio
    A main limitation on the application of thermal therapies is the lack of quantitative temperature knowledge in the treatment zone. This knowledge would increase the safety and effectiveness, through a correct instrumentation control. The portability and relative low cost of the ultrasound devices encourages the application of backscattered ultrasound (BSU) for temperature estimation. So far the methods based on BSU were only developed for homogeneous media. This paper reports a data-driven methodology able to estimate temperature in non-homogeneous media, with the required resolution. Appropriate generalisation performance was observed even under operating situations never considered in the training procedure.
  • Non-invasive time-spatial temperature simulation using neural networks
    Publication . Teixeira, C. A.; Ruano, M. Graça; Ruano, Antonio; Pereira, W. C. A.
    In this paper the performance of radial basis functions neural networks is accessed for non-invasive time-spatial temperature simulation in a gel-based phantom. The medium was heated at different intensities with a physiotherapeutic ultrasound device. In order to find an appropriate neural network structure the multi-objective genetic algorithm was applied. After the structure selection phase a set of preferable individuals was obtained, and the best one presents a maximum absolute error less than 0.5 oC, as desired in hyperthermia. In addition this model has low computational complexity, a fundamental point for a real-time application.