Search for Similarity Transformation Between Image Point Clouds Using Geometric Algebra for Conics

We introduce a novel way of searching for the similarity transformation of pictures using Geometric Algebra for Conics (GAC). In our approach, we do not represent the image objects by their contour but, instead, by the ellipse fitted into the contour points. Such representation makes the consequent similarity search fast and memory-saving and makes the search for the needed transformation easier. Examples of application on the real object images are also included.