Selver,M.A., Ozdemir,M., Selvi,E.
Abstract:
Optical parameter assignment using Transfer Functions (TF) is the sole interactive part in medical visualization via volume rendering. Being an interactive element of the rendering pipeline, TF specification has very important effects on the quality of volume-rendered medical images. However, TF specification should be supported by informative search spaces, interactive data exploration tools and intuitive user interfaces. Due to the trade-off between user control and TF domain complexity, integrating different features into the TF without losing user interaction is a challenging task since both are needed to fulfill the expectations of a physician. By addressing this problem, we introduce a semi-automatic method for initial generation of TFs. The proposed method extends the concept of recently introduced Volume Histogram Stack (VHS), which is a new domain constructed by aligning the histograms of the image slices of a CT and/or MR series. In this study, the VHS concept is extended by allowing the user to define an alignment axis using orthogonal multi planar reconstructions via simple, yet effective, interaction mechanisms. The construction of VHS according to the slices generated specifically for user defined search space allows the more informative integration of local intensity distribution, and better spatial positioning of the organ of interest into the TF. For testing, the proposed strategy is applied to kidney visualization from CT and MRI series. The performance of extended VHS domain is evaluated via intensity based TF design. Volumetric histogram based manual TF specifications are quantitatively compared to VHS based manual tweaking of original slices, and to extended-VHS based automatic TF design. The results show both quantitatively and qualitatively enhanced rendering quality for kidney visualization.