WALSH-FUNCTION CURRENT PATTERNS AND DATA SYNTHESIS FOR ELECTRICAL-IMPEDANCE TOMOGRAPHY

For static image reconstructions in electrical impedance tomography using an iterative algorithm, we may need to collect data from a subject repeatedly. However, in order to reconstruct a static image of a subject at an instant, it is desirable to get all data at that instant. In this paper, we have developed a new data collection method using Walsh functions, which can generate the optimal current patterns for maximal distinguishability and remove the need for repeated data collection. The use of Walsh functions also simplifies the design of current sources since only two levels of current (+1 and -1) are required. Compared to diagonal or neighboring type of pulses as injection current patterns, Walsh injection current patterns provide more information about the interior of the subject since Walsh functions are better at simulating low as well as high spatial frequency patterns. Therefore, Walsh function injection uses the simplicity of pulse type injection and yields the better distinguishability or SNR of sinusoidal injection. Walsh functions also have the advantage of computational simplicity since they take the values of only +1 or -1.