Final Exam Topics
CSE/EE 576 Spring 1995
The final examination will be given on Monday, June 5.
Questions will cover some subset of the following:
1. Additive and "subtractive" color mixing.
2. Segmentation into regions. Definition. Split-and-merge algorithm.
3. shape: convex hulls, distance transforms, medial axis transformations
and digital skeletons.
4. neural nets: perceptrons and perceptron training, linear
separability, fundamental training theorem.
(copies of lecture slides available at Engr. Lib. Copy Center).
5. model matching and pruning using geometric hashing.
In addition you will be asked to describe two of the following
five techniques that were on the midterm exam. You may not choose
the one you already answered on the midterm exam.
-- Edge detection using heuristic search as in Martelli's method.
-- Design of edge-detection masks to maximize signal-to-noise,
ratio, minimize localization error and minimize multiple responses
to the same edge (Canny's approach).
-- Use of the Karhunen-Loeve transformation to efficiently
represent the vectors of a set.
-- Use of Freeman's chain code and Ramer's polygonal approximation
algorithm to obtain good representations of closed contours.
-- Use of the Union-Find abstract data type in an algorithm for
finding the connected components of an image by building a spanning
forest.