TY  - RPRT
UR  - http://www.eledia.org/students-reports/781/
Y1  - 2018///
TI  - Multi?Scaling Bayesian Compressive Sensing Imaging of Dielectric Objects
N2  - In this work, a new Bayesian compressive sensing (BCS)-based imaging technique is proposed to exploit additional information besides that on the target sparsity. More precisely, an innovative iterative multi-scaling (IMSA)-BCS scheme is proposed to combine the a-priori knowledge on the class of scatterers and the progressively acquired information on the location and the size of the unknown object. Accordingly the 2D transverse magnetic (TM) inverse scattering problem is solved by means of an innovative IMSA-based information-driven relevance vector machine (RVM) solver. Some numerical results are shown to verify the effectiveness of the proposed imaging technique.
M1  - technical_report
ID  - elediasc12781
A1  - Anselmi, N.
A1  - Poli, L.
A1  - Oliveri, G.
A1  - Massa, A.
KW  - Born approximation (BA)
KW  -  compressive sensing (CS)
KW  -  inverse scattering (IS)
KW  -  microwave imaging
KW  -  relevance vector machine (RVM)
PB  - ELEDIA Research Center - University of Trento
AV  - public
ER  -