A transputer implemented algorithm for electrical impedance tomography

Abstract

Electrical impedance tomography (EIT) offers a non-invasive technique of great scope for producing clinically important information in a number of medical applications. Initial work by Brown used an isopotential approach to solving the complex backprojection problem. In this study a less restrictive algorithm for EIT has been developed based on the work of Yorkey and of Kaczmarz. Since considerably more processing is required this has been implemented on a transputer in Occam 2 with the aim of achieving real-time imaging. Data were collected, using a Sheffield prototype EIT system, from a test object and a human thorax. Image processing used both the 'Sheffield' and the Yorkey/Kaczmarz algorithms. Our initial results indicate that the images generated using the latter approach were more representative of the source impedance distributions.