Fall 2014 – variational methods in image processing – 048926 

Fall Semester 2014-2015.

Monday 13:30-15:30, room 354.

Course syllabus and info (Hebrew & English)

New course messages

Presentation of project papers will be on Monday, 16.2.2015, at 13:30, room 354.

 

Lectures

Week (date)

Topic

Slides

Notes

1  (20.10) Intro lecture  Intro
1.5 (27.10) Eli Appleboim talk
2  (3.11) Diffusion processes I – linear, Perona-Malik, Weickert  Week 2  Notes 1
3  (10.11) Diffusion processes II - basic properties, stability, numerical implementation.  (no slides)  Notes 2
4  (17.11)  Basic functionals  Week 4  Notes 3
5  (24.11)  BV and Euler-Lagrange  Week 5  Notes 4
6  (1.12)  TV denoising and deconvolution  Week 6  Notes 5
7  (8.12)  Segmentation I  Week 7  Notes 6
8  (22.12)  Segmentation II – level sets  Week 8  Notes 7
9  (29.12)  Optical flow  Week 9  Notes 8
10 (5.1.15)  Convex analysis and spectral TV  Week 10  Notes 9
11 (12.1.15)  Nonlocal operators  Week 11  Notes 10
12 (19.1.15)  Numerical methods  (no slides)Optical Flow
 Notes 11

 

Exercises

Ex#

Topic

PDF

Due date

Files

1 Diffusion  Exercise 1  1.12.2014  Ex1_files
2 TV Denoising and Deconvolution  Exercise 2  22.12.2014  Ex2_files
3 Segmentation and optical flow  Exercise 3  12.1.2015  Ex3_files
4 Nonlocal operators  Exercise 4  26.1.2015  Ex4_files
Proj Final project  Final Proj Presentation 16.2Submission 2.3

Papers for the final project

Paper#

Topic

Authors

PDF

1 Motion detail preserving optical flow estimation  Xu, Jia, Matsushita  Paper1
2 Stereoscopic scene flow computation for 3d motion understanding  Wedel, Brox..  Paper2
3 Consistent binocular depth and scene flow with chained temporal profiles  Hung, Xu, Jia  Paper3
4 Joint estimation of motion, structure and geometry from stereo sequences  Valgaerts, Brhun, Zimmer..  Paper4
5 Robust odometry estimation for RGB-D cameras  Kerl, Sturm, Cremers  Paper5
6 Real-time camera tracking and 3D reconstruction using signed distance functions  Bylow, Sturm, Kerl..  Paper6
7 Real-time non-rigid reconstruction using an RGB-D camera  Zolhfer, Niener, izadi..  Paper7
8 A first-order primal-dual algorithm for convex problems with applications to imaging  Chambolle, Pock  Paper8
9 A fast iterative shrinkage-thresholding algorithm for linear inverse problems  Beck, Teboulle  Paper9
10 The split Bregman method for L1-regularized problems  Goldstein, Osher  Paper10
11 Total generalized variation  Bredies, Kunisch, Pock  Paper11
12 A total variation spectral framework for scale and texture analysis  Gilboa  Paper12

 

 

Older messages

Lecture on 27.10.2014 starts at 14:00, given by Eli Appleboim.