Methods for transforming, or morphing, one image into another have gained in popularity over the past decade. One need only turn on the television set to see this in everything from commercials to music videos. Most of the current commercial work is done using two-dimensional (2D) techniques. Currently needed is method for extending these two-dimensional techniques into three-dimensions (3D). There are many advantages to performing the shape transformation in 3D. One major advantage is the ability to animate the object independently of the transformation. With this added ability one can view the shape transformation from any angle, this is not possible with 2D image transformation. This paper presents a method for computing such 3D transformations. The algorithm uses 3D polygon meshes made up entirely of triangles. Given two polygon mesh models,the algorithm projects them onto a unit sphere and then merges their topological structures. Through the merger, the algorithm produces a new polygon mesh that contains all of the structure of the original two. Transformation can now be performed between the two objects by interpolating between corresponding vertices. Since this algorithm uses the topological and geometric data of both objects, the resulting shape transformation exhibits less distortion than other algorithms.

Sample Morph Mpeg