An improved back-projection algorithm for magnetic induction tomography based on magnetic field lines

Magnetic induction tomography (MIT) is a non-invasive and non-contact medical imaging technology. It is a visualized implementation based on image reconstruction algorithms, which plays an important role in determining the image quality. In this paper, an improved back-projection algorithm based on magnetic field lines is presented and can be described as follows. Firstly, the phase data collected at the imaging region boundary coil is normalized. Secondly, the back-projection matrix is simply weighted based on the difference of back-projection paths and the sensitivity to conductivity change of the pixel in the imaging region. Finally, a data modification model is proposed based on the electromagnetic relation in MIT. The results show that the improved algorithm can accurately reflect the distribution of conductivity and the object position in the imaging region with a high resolution. Specifically, when the object is located at the boundary of the imaging region, the back-projection algorithm of MIT can improve the convergence of the edge of the detected object. The anti-noise performance of improved algorithm could meet the standard of applications. In a measurement system with SNR is over 20dB, images could be successfully reconstructed without being affected by the noise.