Abstracto

Early detection of breast cancer using signal calibration neural network (SCNN) technique

S Yuvarani, Jothi Venkateswaran

In this work, a wearable clinical prototype with patient interface for microwave breast cancer detection is designed. It operates in the 2-4 GHz range and contains 16 wideband sensors embedded in a hemispherical dielectric radome. The cancer cell of size below 5 mm are detected and measured from breast phantom. The subsequent scan acts as a calibration signal and the skin reflections will be suppressed if the skin contour remains consistent throughout the rotation. However, the scattering response from within the breast also becomes distorted in the process as twin targets are often reconstructed from single scatterers and significant targets located near the axis of rotation are often eliminated. The proposed work designed to perform multistatic signal calibration with minimal distortion of internal breast scatterers; The proposed work offers superior tumor identification, accurate localization, and strong artifact resistance over existing wavelet algorithms. In this paper we investigate how signal processing can be accelerated for diagnosis by using NN. The various scenarios: homogenous and heterogeneous breast models with varied densities, combining both ideal and practical signal analysis methods were taken. Extensive simulations and analyses using backscattered signals received from wearable breast models were conducted to validate the performance of the proposed algorithm. From the results, we can measure the cell with 95.6% in accuracy.