Image processing in the digital tomosynthesis for pulmonary imaging

Abstract

Digital tomosynthesis makes it possible to reconstruct multiple tomographs from digital data obtained during a single tomographic motion and permits digital processing, which adds a number of special advantages to the well-known advantages of conventional tomography. We performed digital tomosynthesis with a fluororadiographic TV unit with tomographic function which was capable of producing pulsed low-and high-energy X-rays alternately, and we studied digital image processing to improve the image clarity of the reconstructed tomographs. To identify the optimal parameters for processing image data by means of spatial frequency filtration we evaluated the spatial frequency distribution of image data in linear tomographs of the lung, and on the basis of the results of this study we developed several types of digital image processing to reduce tomographic blur and system noise, to improve visualisation of faint opacities, to reduce resistant tomographic blur as well as overall blur, and to generate low-noise bone images based on dual-energy subtraction tomosynthesis.