Vertical Interior Distance Ratio to Minimum Bounding Rectangle of A Shape

This paper proposed a simple shape descriptor, based on the vertical interior distance ratio of the minimum bounding rectangle (VIDR). Shape descriptor is widely used to describe shape, especially in leaves matching and trademark retrieval. VIDR is the proportional distribution of the vertical interior distance between the shape contour points and its four sides of the minimum bounding rectangle. The minimum bounding rectangle can change according to the change of shape scale and direction, which is extremely suitable for describing the changing shape, and it has global characteristics. Compared with the descriptor of centroid contour distance (CCD) and the shape context descriptor (SC), which are all use contour points to describe shape, our descriptor has a vertical direction and is able to distinguish the same centroid shapes or similar contour shapes, it also has simpler feature extracting description process. More importantly, the experimental results show that our descriptor has a higher precision and faster speed.