The topic of this paper includes graph cuts based computed tomography reconstruction methods in binary and multi-gray level cases. This approach combines the graph cuts and a gradient based method. The present paper introduces and analyses the shape circularity as a new regularization and incorporates it in a graph cuts based computed tomography reconstruction approach, thus introducing a new energy-minimization based reconstruction algorithm for binary tomography. Proposed method is capable for reconstructions in cases of limited projection view availability. Results of experimental evaluation of the considered graph cuts type reconstruction methods for both binary and multi-level tomography are presented.

本文的主题包括在二元和多灰度级情况下基于图割的计算机断层扫描重建方法。这种方法结合了图割和基于梯度的方法。本文介绍并分析了形状圆度作为一种新的正则化方法，并将其结合到基于图割的计算机断层摄影重建方法中，从而介绍了一种基于能量最小化的新型二元断层摄影重建算法。所提出的方法能够在投影视图可用性有限的情况下进行重建。提出了针对二值和多层层析成像技术的图割型重构方法的实验评估结果。