CS 482 Midterm (100p), 10/24/2002
Name:
1. (15p) You are given a camera with adjustable focal length. Describe how it could be used for
size constancy; i.e., to make different objects appear to have the same image sizes. Show the
relevant math.
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
Guidelines and tips
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
Computer Vision - Midterm Exam 2002 | CS 482, Exams of Computer Science
Material Type: Exam; Professor: Duric; Class: Computer Vision; Subject: Computer Science; University: George Mason University; Term: Unknown 1989;
Typology: Exams
Pre 2010
1 / 5
This page cannot be seen from the preview
Don't miss anything!
Related documents
Partial preview of the text
Download Computer Vision - Midterm Exam 2002 | CS 482 and more Exams Computer Science in PDF only on Docsity!
CS 482 Midterm (100p), 10/24/
Name:
- (15p) You are given a camera with adjustable focal length. Describe how it could be used for size constancy; i.e., to make different objects appear to have the same image sizes. Show the relevant math.
- (30) You are given the following binary image.
Assume that a coordinate system with x-axis pointing right and a y-axis pointing up is added. The coordinates of the left bottom pixel are (0, 0). Compute the following parameters:
a) (3p) area, b) (3p) perimeter (describe your algorithm), c) (4p) center of mass, d) (10p) second order moments M 11 , M 20 , and M 02. e) (10p) central moments μ 11 , μ 20 , and μ 02.
- (50p) You are given a template N and a binary edge image M.
a) (5p) Represent N so that it can be used by the GHT (Generalized Hough Transform) algorithm. b) (15p) Use GHT to find N in M. How many matches did you find? You can use the grid above for the GHT accumulator array. c) (10p) Use forward/backward pass algorithm to compute the distances of all points in M from the edge points of M. You can assume that the distance between any pair of 8-neighbors is 1. d) (15p) Use Chamfer matching to find N in M. Show all placements with scores lower than 10. You can assume that the placement of N is at the position of the topmost pixel of N. What are the best matches? e) (5p) Show how the distance transform array could be used to perform a Hausdorff dis- tance matching on N and M. Show the best match(es).
Template N, edge image M, and the GHT scores.
N M
N and M GHT scores
Distance transform and Chamfer match scores
forward pass backward pass match score
Hausdorff distance matching