ECE 661 Computer Vision (Fall 2006)

---

For the latest syllabus for the course,   click here.

---

Instructor: Avi Kak

Teaching Assistant: Pradit Mittrapiyanuruk

Email: mitrapiy@purdue.edu

Office : EE180 Phone: 49-43456

Office hours:  Mon-Fri 11 am-5 pm. (not available on Wed 1-3 pm.)

---

Homework 1

Problem

Sample Solution (Courtesy of Kihyun Hong)

Sample Solutions for Homework 1 &2 (Courtesy of Yufeng Li)

Sample images(Courtesy of Akio Kosaka)

Homework 2

Problem

Sample Solution (Courtesy of Kihyun Hong)

Sample Solutions for Homework 1 &2 (Courtesy of Yufeng Li)

Homework 3

Problem

Sample Solution (Courtesy of Dave Kim)

Homework 4

Problem

Sample Solution (Courtesy of Dave Kim)

Sample Solution (Courtesy of Kihyun Hong)

Homework 5

Problem

Sample solution (Courtesy of Kihyun Hong)

Sample solution (Courtesy of Wonjo Jung)

PDF File of the calibration pattern (courtesy of Hyukseong Kwon)

Sample Images taken in our lab

Note that: a rough value of the intrinsic parameter (obtained from our implementation) of the camera that captures these images is

K =[ 732.9653 0.7232 320.0412

0 731.1049 231.5993

0 0 1.0000 ];

Handout: A Memo on How to Use the Levenberg-Marquardt Algorithm for Refining Camera Calibration Parameters (Latest modified on Nov 14, 2006)

Sample MATLAB script (incomplete): the Levenberg-Marquardt Algorithm for Refining Camera Calibration Parameters (Latest modified on Nov 14, 2006)

Sample MATLAB script: Symbolic computation of the Jacobian

Notes

1. Dr.Jonhny Park has suggested a C/C++ implementation of Levenberg-Marquardt algorithm for minimizing an objective function at the following URL:

http://www.ics.forth.gr/~lourakis/levmar/

           2. For people who are using the function "cvHoughLines2" in OpenCV, there are some minor error in the documentation. Please see the below links for the correction. Thanks Paul Shin and Samuel Kim for this important information.

· cvHoughLines2 Docu Sample error (from Yahoo OpenCV Community posted by frankoni)

· Similar sample code by Samuel Kim

3. For students who do not want to use OpenCV in their homework, here are the links to some sample C codes: for both Canny Edge Detector and Hough Transform

· Canny Edge Detection: The implementation of Mike Heath @ University of South Florida

· Hough Transform: The implementation from the book “Practical Algorithms for Image Analysis” M.Suel at el

Homework 6

Problem

· A sample MATLAB script for Algorithm 11.1 by Peter Kovesi

· A sample MATLAB script for Image Rectification by Du hyun (main page)

- Handout by Hagen Spies (p.12-19) summarizes the idea of image rectification. Thanks to Sameul Kim for the reference.

Sample solution (Courtesy of Wonjo Jung)

Sample solution (Courtesy of Yunfeng Li)

Sample solution (Courtesy of Kihyun Hong)

Homework 7

Problem (Optional)

· Pollefeys, Koch and Van Gool IJCV 1999 paper

Sample implementation

(Note: This implementation is based on the linear approach presented in section 4.2 of Pollefeys 99 paper. Its correctness is not confirmed. Also it is not quite complete yet.)

Exam 1

the exam

Exam 2

the exam

---