Learning Objectives for SIGGRAPH 2004 Project Papers

Fragment Based Image Completion

• Image Compositing using Laplacian and Gaussian Pyramids
• The Push-Pull Algorithm for Scattered Data Interpolation
• Searching for matching fragments (including similarity measurement and adaptive region sizes)
• Handling of confidence values and level sets
• Analysis of the Fragment Based Image Completion technique

View Dependent Displacement Mapping

• Basics of bump mapping, horizon mapping, displacement mapping
• Pixel shader ops and their displacement in the graphics pipeline. How does VDM fit in?
• VDM texture generation (issues include terrain rendering and handling curvature)
• Rendering with VDM textures (including handling of self shadowing)
• How eigendecomposition/SVD is used in compression
• How to prepare a matrix so that it compresses well
• Analysis of the VDM technique

Graphcut Textures

• The basics of texture synthesis; previous techniques such as Efros and Freeman and Wei and Levoy and how they work and what are their strengths and weaknesses
• Maxflow/Mincut Algorithm and how it applies to texture synthesis
• Patch placement and matching (including cost functions and rapid computation)
• Accounting for old seams
• Feathering
• Analysis of the Graphcut Textures technque

Boolean Operations on Surfel Bounded Solids

• The surfels approach -- Why they are used, how they are used, their strengths and weaknesses.
• Intersection testing within a leaf cell - and associated problems.
• The resampling operator - Why it's necessary and how and why it works.
• Octrees and how they work.
• Basic approaches to finding nearest neighbours efficiently
• Analysis of the surfels technique