[Next] [Up] [Previous]
Next: About this document Up: Computer Vision Previous: A POP-11 Procedure to

Exercises

  1. List as many different kinds of use for human vision as you can (e.g., looking for something, noticing the sudden appearance of an object, planning a route down a mountain).
  2. Design a model for the appearance of an object in your immediate environment. You will need to force yourself to think about the object in terms of its 2-dimensional projection on your retina rather than as a 3-dimensional entity. Express your model as a semantic net of the form shown in figure 8.11.
  3. Write a program to turn an image into its negative version. You will need to replace each grey level by the difference between that level and the level representing maximum intensity (99 in our case).
  4. Outline a set of rules for finding closed edge contours in an image by hopping between neighbouring edge elements. If you are feeling ambitious, turn this outline into a program.
  5. Think about how you might set about blurring an image and write a program to achieve this. (Hint: You will need to replace each grey-level with a new value computed from grey-levels at neighbouring pixels.)
  6. One way in which to model the 3-D structure of objects is in terms of a number of primitive shapes glued together. For example, we can represent the limbs and body of many animals using cylinders (see Marr, 1982). Try to draw up a list of primitive shapes with which to model the objects in a typical office scene. For example, you could use a series of rectangular faced shapes to construct the tables.


Cogsweb Project: luisgh@cogs.susx.ac.uk