A new recognition method for partial 2-D perspective shapes

The authors present a new method for the recognition of partial two-dimensional (2-D) perspective shapes. The method can identify the unknown shape directly from one image taken by a camera from an arbitrary perspective. The approach derives the geometrical relationship between the image frame and the object frame, develops the mapping formula to recover the shape from the perspective projection, and recognizes the recovered shape by estimating the fast shape descriptor coefficients and the parameter perimeter²/area. In the method, the number of missing boundary points need not be known. The method is simple and fast. Experimental results from several 2-D shapes indicate that a satisfactory recognition rate can be obtained even when 20% of the boundary points are missing