About

About Me

I am a third year undergraduate student reading Computer Science at The University of Sheffield, UK. As part of my course I am required to undertake a personal project which accounts for half the year's overall mark.

The dissertation project involves research in a given area. From this initial direction students can expand the project in a manner of their choosing, as agreed with their Project Supervisor. The projects consist of literature reviews, individual work (such as programming), a poster session (in which a visual poster is displayed, any software demonstrated and questions answered), possible vivas, and a final bound document.

Morphing

Morphing is an animation and effects technique that has become popular in films, adverts, music videos and scientific visualisations - among many other applications - for more than a decade. It has the effect of appearing turning one object into another in a natural and fluid manner. This is typically achieved by creating a warp animation for both keyframes, so that the two objects' shapes are identical in corresponding frames, changing from the source object's shape to the destination object's shape as the animation progresses. This shape correspondence results in pixel correspondence, so to complete the morph all that needs to be done is a pixel-level dissolve, such as a cross-fade, run from one animation to the other.

The most time consuming part of morphing is the specification of the shape. This is most often done using a mesh or a set of features to mark out identical points on both images, from which the warps are generated. Since these features determine the objects' shapes they are vital to the quality of the morph.

About the Project

The idea of a morphing project was proposed by Dr. Steve Maddock in 2002 with the aim of creating a program that could either produce advances in the area of automatic morphing (reducing the work of animators specifying corresponding points) or look into hierarchical morphing (in which fine details can be animated independently of larger forms).

Unfortunately I was unable to take the project that year, but Dr. Maddock kindly reserved it for me in 2003. During the course of that year the project specification changed. While the morphing aspect remained, the actual aim was to create a morphing program which had a real-world use, specifically in the domain of Psychological research being studied by Dr. Olivier Pascalis in the University's Psychology Department. He plans to use the software to allow one of his undergraduate project students conduct follow-up research into how people prototype human and animal faces by using a three-way morphing sequence to determine at what point a morphed pseudo-hybrid of the faces of two animals stops being one and becomes another.

• Initial project description
• Revised project description

The Psychology Department intends to collect data early in 2004, so requires a working system before then. It is my intention to use work I did over the Summers of 2002 and 2003 to release a basic morphing program early in the first semester of 2003. This leaves plenty of time to alter the system to the mutual satisfaction of all parties. I can then build further on the system, particularly in areas relating to the user's specification of the morph. I will look at topics such as semi-automated morph specification including edge detection and templates. Any future version should add improvements that are of real benefit to the Psychology Department in any repeated, related or follow-up work.

The final project deadline will be in May 2004.