SCM-5: Auto-Morphing and Hierarchical Morphing

Supervisor: Steve Maddock

Most morphing techniques rely on the user identifying key parts of the two images to be morphed. Typically such parts then have equal weight when determining the morph. The aim of this project is to investigate one of two ideas. The first idea is to investigate techniques to automate the whole morphing process. Here, edge detection algorithms will be investigated to see if it is possible to automatically identify key parts and match them in pairs of images. The second idea is to look at hierarchical morphing techniques where the morphing of general and specific features can be separately controlled, e.g. the general shape of a face versus specific details on the face. Here a hierachical ellipse-based morphing techniques is proposed and this can be compared with the common point or line-based techniques.

Prerequisites: Good maths and programming ability

Resources: Java or C++

Initial reading: Beier, T. and Neely, S, "Feature-based image metamorphosis", Computer Graphics, 26 (2), July 1992, pp.35-42