抄録
With the developments of modern high-speed digital computer, X-ray as well as MRI (Magnetic Resonance Imaging) tomography is widely used as one of the most effective instruments for medical diagnosis. On the other side, EIT (Electrical Impedance Tomography) is developing only for the industrial use. This is because EIT is one of the functional types of tomography, which require a solution of ill posed system of equations while the X-ray and MRI need not solve such an ill posed system of equations.
In order to overcome such difficulty in EIT, this paper proposes a new deterministic method, which consists of a new solving strategy, i.e. GVSPM (Generalized Sampled Pattern Matching) for the ill posed linear system of equations and image processing tool, i.e., convolution strategy to extract the common parts in the images. As a result, it is revealed that our method promises the highly reliable tomography.
