Electrical impedance tomography based on BP neural network and improved PSO

A new method for static electrical impedance tomography was proposed in this paper. The new algorithm was based on the weight adjustments of error back propagation of BP neural network whose weights and thresholds were modified by improved particle swarm optimization. This method can not only well adapt to non-linear and ill-posed characteristics of electrical impedance tomography, but also overcome the limitations both the slow convergence and the local extreme values by basic BP algorithm. The improved particle swarm optimization has less iteration and higher accuracy then the standard particle swarm optimization. Experimental results show that the method is easy, fast and can effectively improve the image resolution.

[1]  David Isaacson,et al.  Electrical Impedance Tomography , 1999, SIAM Rev..

[2]  Gary B. Lamont,et al.  Visualizing particle swarm optimization - Gaussian particle swarm optimization , 2003, Proceedings of the 2003 IEEE Swarm Intelligence Symposium. SIS'03 (Cat. No.03EX706).

[3]  David R. S. Cumming,et al.  Electrical impedance tomography for sensing with integrated microelectrodes on a CMOS microchip , 2007 .

[4]  Michael N. Vrahatis,et al.  Recent approaches to global optimization problems through Particle Swarm Optimization , 2002, Natural Computing.

[5]  James Kennedy,et al.  Particle swarm optimization , 1995, Proceedings of ICNN'95 - International Conference on Neural Networks.

[6]  Tadakuni Murai,et al.  Electrical Impedance Computed Tomography Based on a Finite Element Model , 1985, IEEE Transactions on Biomedical Engineering.

[7]  Brian H. Brown,et al.  Imaging spatial distributions of resistivity using applied potential tomography , 1983 .