Modeling, Measuring and Perceiving Appearance



The appearance of the everyday world is influenced by many factors such as shape, color, texture, reflection properties and the surrounding environment. The causes of appearance have long been studied by scientists, and of interest to almost anyone. Recently modeling and simulating appearance has been studied by researchers in computer graphics and computer vision. By modeling the appearance of objects and materials, we are able to render virtual versions of those objects. By measuring appearance, we are able to digitize real world objects and place them in new synthetic environments. Measurement techniques also allow us to experimentally validate different appearance models. In computer vision, texture and reflection models can be used to recognize scenes and objects in images.

Here is some inspiration for the types of appearance we will be discussing (diffuse, glossy, specular, and anisotropic surfaces, hair, flowers, objects with and without subsurface scattering).

The class is open to students with a background in computer graphics or computer vision. The class may be taken for 1 or 3 credits. For 1 credit, each student will be expected to participate in all class activities; for 3 credits, a final project is also required.


Gates Room 392, Tuesdays and Thursdays from 2:30-3:45pm

Technical Lectures

  1. Preliminaries (Hanrahan)
  2. Subsurface Scattering I: Kubelka-Munk model (Hanrahan)
  3. Subsurface Scattering II: Hanrahan-Krueger model (Hanrahan)
  4. Subsurface Scattering III: dipole BSSRDF model (Hanrahan)
  5. Scattering from Fibers (Hanrahan)
  6. Sensing for Appearance Acquisition (Lensch)
  7. Slices of the 8D Reflectance Field (Lensch)
  8. Model Fitting and Analysis (Lensch)
  9. Data-driven Models and Representations (Lensch)
  10. Psychophysics of material perception (Hanrahan)
  11. Recognizing materials (Hanrahan)