Potential optimal designs for visualization techniques of multidimensional data can be as varying as the set of tasks to be fulfilled by it and possible use cases combined. For this reason visualization techniques in Scientific Visualization (SV) need to be designed task specifically. In case of this work, the goal was to design a visualization encoding of similarity data in the 3D spatial domain with three attributes at each position. To do so, tasks to classify requirements that would need to be fulfilled were derived. This thesis details the corresponding step of designing and then implementing the decided on visualization solution. This obtained visualization is constituted by three major visual encoding elements: Segment Borders which are surfaces of segment boundaries extracted from the similarity volume by means of a spatial partitioning technique, Segment Glyphs which are glyphs that encode the similarity values and Gradient Glyphs which are glyphs that encode the gradient of the similarity data. The applicability of the determined visualization was examined in usage scenarios. Additionally, the approach was evaluated in an expert interview. While the gradient glyphs were regarded skeptically the general approach for the segment visualization was seen as a good advance to visualize similarity data in the 3D domain with the particular advantages the domain presents.