Efficient techniques to enhance nearfield imaging of human head for anomaly detection

Abstract

This paper proposes efficient algorithms to enhance the nearfield electromagnetic imaging of human head. Forward problem is modeled using SAM head phantom with brain tumor anomalies, surrounded by a circular applicator antenna array. Scattered signals are compressively sensed (CS) at a limited number of sensing positions, and the sensed signals are preprocessed efficiently using a proposed novel technique to maximize information extraction. A dictionary is formed and then implemented in CS based inverse problem analysis. Reconstructed images are enhanced using new post-processing techniques to improve the spatial resolution. Image quality is analyzed using the quality metric in terms of peak signal-to-noise ratio (PSNR). The quality of the reconstructed images and the corresponding PSNR values reveals the validity of the imaging techniques.