Reducing the complexity of scanned meshes while maintaining the shape indicated by the data points as well as possible.

Paper Presentations by Balaji Kannan and Satish Kumar

Michael Garland and Paul Heckbert: "Surface Simplification Using Quadratic Error Metrics" SIGGRAPH'97

• What is the basic simplification step ?
• What may be the problem when this algorithm is used in solid modeling or rapid prototyping applications ?

Hugues Hoppe: "Progressive Meshes" SIGGRAPH'96

• In part 2 of the paper, what is the criterion that decides which edge should be collapsed next ?
• What tests are made to make sure that a proposed edge collapse does not damage ?

META-SKILL:
How to read a paper if you do not have enough time.

• Find out what the key idea is.
• Where in the paper is the key idea presented in detail.
• Try to understand that section.

• -- Read abstract.
• -- Skim introduction to where new idea is discussed.
• -- Read the paragraph(s) on new idea carefully.
• -- Skip history, previous work.
• -- Look at figures, study figure captions.
• -- Study key section(s) on new idea.
• -- Skip detailed mathematics proofs.
• -- Skim result section to get a feeling of how much the idea brought.
• -- Skim discussion and conclusions to understand caveats, problems.
• -- Ignore appendices.
• -- Write down your key insights:
  • - what is the key idea ?
  • - what are the critical algorithmic steps ?
  • - what are some of the important new quantities introduced ?
  • - how are they calculated ?
  • - how are they used ?
  • - what has been achieved, demonstrated, proven ?
  • - how well does the new idea work ?
  • - where in your own work might it be relevant ?

Assignment: READ the papers for the upcoming lectures and answer the stated questions