︎

V1.1: To transfer the information from the sphere to the plane the method of dimensionality :reduction can be applied. (Please use arrows to navigate through the slides.) V1.2: Through the geometrical operation of projection all points of the 3D object are transferred orthogonally to the plane. V1.4: Let’s analyze three points according to their distance relations. We project them onto the plane like before. V1.5: But when we measure the distances between them a problem occurs: the relations of the distances for the same points differ considering the sphere and the plane. Due to the reduction of dimensionality the relations get distorted, and the results are less descriptive. V2.1: There is another method to display the content in two-dimensional form which does not rely on reduction. V2.2: Suppose we determine a point on the sphere and try to find the nearest neighbors. With a simple mathematical operation, the other points of the sphere can be arranged around the selected point according to their similarity/proximity. These relation can then be plotted in 2D. V2.3: The same operation is then performed for other points. Successively, the multitude of singular individual representations create a scan of the original sphere (similar to the 3D-scanning method of photogrammetry in which 3D objects are extracted from 2D photos. The more pictures from different angles, the more detailed the resulting 3D-scan). Contrary to the example of V1, this methods produces a induvial “map” for each single perspective.