4-D fuzzy connectedness-based medical image segmentation technique

In this paper a new 4-D fuzzy connectedness approach to medical image segmentation is presented. The developed algorithm dedicated to the analysis of magnetic resonance studies, improves the segmentation results by introducing simultaneous analysis of different MR projections. It is applicable especially for those cases, in which the slice thickness or slice gap are relatively big, but different sequences are available. The proposed modification to the fuzzy connectedness analysis introduces new adjacency model, taking into consideration spatial conditions of two or more acquired projections. The developed adjacency model uses the spatial series connections, generated on the DICOM header base. The presented methodology is evaluated on the database of bone tumours images.