Title :
Image classification for a time-domain microwave radar system: Experiments with stable modular breast phantoms
Author :
Santorelli, Adam ; Laforest, Olivier ; Porter, Emily ; Popovic, Milica
Author_Institution :
Dept. of Electr. Eng., McGill Univ., Montreal, QC, Canada
Abstract :
In this paper, we report on durable, long-lasting rubber breast phantoms that contain four tissue types (skin, gland, fat, and tumor) that accurately mimic their dielectric properties. An improved fabrication procedure for creating stable high-water content tissue phantoms is demonstrated. The phantoms are made in a modular fashion such that the configuration, including tumor depth, size, and presence of tumor, can be varied as desired. Further, we reconstruct images of the phantoms using our time-domain microwave radar system. The images are then used to test an image classification algorithm that aims to identify if the image under consideration is one of a healthy breast or a breast that contains a malignancy.
Keywords :
biological organs; image reconstruction; medical disorders; medical image processing; microwave imaging; phantoms; radar imaging; skin; tumours; dielectric properties; fat; gland; high-water content tissue phantoms; image classification algorithm; image reconstruction; malignancy; modular fashion; rubber breast phantoms; skin; stable modular breast phantoms; time-domain microwave radar system; tumor; Breast; Dielectrics; Imaging phantoms; Permittivity; Phantoms; Plugs; Tumors; cancer detection; image classification; microwave imaging; phantoms;
Conference_Titel :
Antennas and Propagation (EuCAP), 2015 9th European Conference on
Conference_Location :
Lisbon