Automatic Sensor Reconfiguration based on Adaptive Relevance Vector Machine for Uncertainty Reduction in Tomography Imaging

We apply the Adaptive Relevance Vector Machine to automatically select the measurement set in a tomographic setting, from all the arrangements or combinations of the measuring elements, that yield the lowest level of uncertainty about the estimated results, while maintaining good image reconstruction. To illustrate the proposed method, we present simulation results derived from Electrical Capacitance Tomography.