Electrical Impedance Tomography Image Reconstruction with apriori Knowledge for Gesture Recognition

Electrical impedance tomography (EIT) reconstructs the conductivity images of an object based on the voltages and the currents at the surface of the object. Among the clinical applications of EIT thorax imaging, brain function monitoring and breast cancer screening. However, EIT suffers from a poor signal-to-noise ratio and spatial resolution. To improve the quality of reconstructed images and provide clear information about the structure, the paper proposes to include prior knowledge about the structure from MRI. The proposed techniques could be very promising with more work. To illustrate this technique in this paper, eight electrodes with the adjacent current pattern were employed in this research. For each gesture, forty data points are recorded. The Gauss-Newton and Back-Projection algorithms are used for image reconstruction with prior knowledge of the human forearm from MRI images. The resultant contours of the Forearm, Ulna, and Radius have been used to build the mesh. In the Gauss-Newton algorithm, the results reveal that structural information improves image quality and conductivity distribution, particularly in the region of two bones. In addition, the reconstructed images using the Gauss-Newton algorithm performed better than images reconstructed using the back-projection algorithm.