# CMSC/AMSC 498D Spring 2012: Deblurring Digital Images

Taught by Dianne P. O'Leary

oleary@cs.umd.edu

When and Where: Fridays ...... 11-11:50am (CSI 1122)

Resources:

• Overview of the course
• Course syllabus
• Chapters from the textbook and other material of interest
• Answers to diagnostic quiz. 1 class participation point for turning it in; 2 for great performance. (Remember that there are 120 possible class participation points, so this will not make or break your grade!)
• Answers to first participation exercise on 02-03.
• Answers to first participation exercise on 02-10.
• Answers to first participation exercise on 02-17. For the last part of the 2nd exercise, there are several alternatives:
• Use a double for loop to perform the addition.
• Add shifts of the image. For example, to average each pixel of P with the one to its right:
Pa = P;
Pa(:,1:end-1) = .5 * ( Pa(:,1:end-1) + Pa(:,2:end) );
• Use circshift, but this does does odd things for the border pixels.
• Easiest: Use imfilter.
• The textbook's website, where you can find the software, the data for the challenges, and additional challenges and data.
• New! 2/14/2012 If you think a mistake has been made in grading your work, submit it for regrading within two weeks of the date on which the work was returned to the class. After that, the grade will be considered final. Keep your work in case there is a question about recording of grades.
• GRACE: These are the computers that give all of you access to Matlab. (If you have a more convenient option, it is fine to use it.)
• Accessing Matlab on the GRACE machines, with graphics. Helpful summary of things to know, from a student.
• Information about computer accounts. See also the additional pointers at the bottom of notes by Larry Herman. For your work, you may use any machine that has the necessary software.
• Sources for Matlab information:
• Official Matlab documentation
• Matlab Primer: 39 pages of basic information
• Timothy A. Davis, Kermit Sigmon, Matlab Primer, CRC Press 2005. A 200 page version of the above reference.
• D. J. Higham and N. J. Higham, Matlab Guide, SIAM Press 2005.
• Lecture notes:
• Introduction and notes for Chapter 1.
• Notes for Chapter 2.
• Notes for Chapter 3. (From Per Christian Hansen)
• Notes for Chapter 5. (From Per Christian Hansen)
• Notes for Chapter 6.
• Demonstrations:
• chapter2demo.tar
• SCCSproj1solution.m and proj1data.mat
• gcv.m and gcvfun.m
• discreptik.m
• discreptsvd.m
• Schedule for Spring 2012 New! 03/27/2012 Now includes participation point values for each class and the term project presentation schedule.

 01 3 pts Jan 27 Intro., Course Organization, Diagnostic Quiz Preparation: None necessary. 02 6 Feb 3 Cameras, deblurring, and matrices. Preparation: Download Challenges.zip and HNO.zip from book's website. Work through Chapter 1. 03 8 Feb 10 Images in Matlab. Preparation: Work through Chapter 2. 04 8 Feb 17 Images and Deblurring Preparation: Work through Chapter 1 and 2. Bring a laptop if you can. 1st class participation: Frobenius norm. 05 8 Feb 24 Deblurring with the SVD. Preparation: Work though Chapter 3. 06 8 Mar 2 A GUI for deblurring. Preparation: None. 07 8 Mar 9 Regularization by Spectral Filtering. Preparation: Work through Chapter 5. 08 6 Mar 16 Regularization by Spectral Filtering. Preparation: Work through Chapter 6, Sections 1-4. Mar 23 Enjoy Spring Break! Preparation: Relax ... 09 6 Mar 30 Choosing regularization parameters. Preparation: Work through Chapter 6, Sections 4-6. 10 5 Apr 6 Choosing regularization parameters. Preparation: Bring a laptop if you can. 11 9 Apr 13 3 term projects: H, C, R. 12 9 Apr 20 3 term projects: V, Z, U. 13 5 Apr 27 Activity Preparation: Bring a laptop if you can. 14 9 May 4 3 term projects: I, G, P. 12 May 11 4 term projects: D, F, W, T. Optional meeting, 11am - 12:30pm. 15 12 May 15 6 term projects: M, Y, S, Q, H, C. (Final exam period) Note special time: Tuesday, 8:30-10am.