M2. Optimization and Inference techniques in Computer Vision (31819)

The aim of this module is to learn about numerical optimization algorithms that are behind many tasks in computer vision. Main concepts will include energy minimization, numerical techniques for variational problems, convex optimization and graphical models. These techniques will be applied in the project in the context of image segmentation and restoration (denoising and inpainting).
Project title: 
Removing objects in urban scenes
The goal of this project is to gain practical experience with the basic optimization methods used in computer vision. The methods will be studied in the context of image segmentation and restoration (denoising and inpainting), with special emphasis on the formulation of the optimization problem and its resolution. The tools learnt along this project are generic and present in a majority of computer vision applications (as found in other modules, e.g. in clutter removal for improved 3D scene reconstruction, M6).
Module lectures: 

 

Academic Year 2016-2017        
Week Date Time Lecture Lecturer University Building Room
1 Tue. Oct. 4th  16:00 -18:00 Introduction to energy minimization methods. Overview of varational formulation. Examples.

Juan Fco. Garamendi (1h)

Coloma Ballester (1h)

UPF  Roc Boronat    52.221 
Numerical techniques for variational problems (Examples)
Thu. Oct. 6th 16:00 -18:00 Review of numerical linear algebra: minimum squares, regression, singular, value decomposition, interative methods,applications.  Coloma Ballester UPF Roc Boronat  52.S31
Thu. Oct. 6th 18:00 - 19:00 Project Introduction

Juan Fco. Garamendi

Vadim Fedorov

UPF  Roc Boronat

(2 rooms)

52.201

52.213

               
2 Tue. Oct. 11th 16:00 -18:00 Euler-Lagrange equation. Gradient descent and advanced numerical schemesfor partial differential equations (Multigrid). Applications and examples (Inpainting and total variation restoration). Juan Fco. Garamendi UPF  Roc Boronat  52.023
Thu. Oct. 13th 16:00 -18:00 Convex optimization. Constrained and unconstrained optimization. Primal, dual, and primal dual methods. Convex relaxation. Applications (Total Variation restoration, disparity,optical flow computation, minimization of non-local functionals). Coloma Ballester UPF  Roc Boronat  52.S31
Thu. Oct. 13th 18:00 - 19:00 Project follow-up

Juan Fco. Garamendi

Vadim Fedorov

UPF  Roc Boronat
(2 rooms)

 52.425

 52.427

               
 3 Tue. Oct. 18th  16:00 -18:00

Extension to multi-channnel images and more-than-two dimensional images Applications and examples (Munfurd-Shah model, medical image , video editing, optical flow). 

 Juan Fco. Garamendi  UPF  Roc Boronat  52.S31
Thu. Oct. 20th  16:00 -19:00 HOMEWORK        
               
4 Tue. Oct. 25th 16:00 -18:00 Bayesian networks and MRFs. Inference . Example: stereo matching.  Joan Serrat UPF  Roc Boronat  52.023
Thu. Oct. 27th 16:00 - 18:00 Inference algorithms 1: belief propagation (sum-product and max-sum) and generalizations. Examples: bob tracking. Oriol Ramos UPF  Roc Boronat  52.S31
Thu. Oct. 27th 18:00 - 19:00 Project follow-up

Juan Fco. Garamendi

Vadim Fedorov

UPF  Roc Boronat    (2 rooms)

52.425

52.427

               


5

Tue. Nov. 1st   HOLIDAY        
Thu. Nov. 3rd 16:00 -18:00 Inference algorithms 2 : Graph cuts, linear programming relaxation. Oriol Ramos UPF  Roc Boronat 52.S31
Thu. Nov. 3rd  18:00 - 19:00 Project follow-up

Juan Fco. Garamendi

Vadim Fedorov

UPF  Roc Boronat

52.425

52.427 

               
6 Tue. Nov.8th  16:00-18:00 Learning of graphical models.Structured SVMs Joan Serrat UPF Roc Boronat 52.S27
Thu. Nov.10h 16:00-18:00 GM for computer vision : two guided excercises .

Oriol Ramos 

Joan Serrat

UPF TALLERS 

(2 rooms)

   54.004

   54.005

Thu. Nov.10th 18:00 -19:00 Project follow-up

Juan Fco. Garamendi

Vadim Fedorov

UPF Roc Boronat

 (2 rooms)

   52.425

   52.427

               
7 Thu. Nov.17th 16:00-19:00 Project Presentations 

Juan Fco. Garamendi

Vadim Fedorov

UPF Roc Boronat

 (2 rooms)

   52.329

   52.S31

               
8 Tue. Nov. 22nd  16:00-19:00 HOMEWORK        
Thu. Nov. 24th  16:00-19:00 HOMEWORK        
               
9 Thu. Dec. 1st 16:00 -19:00 EXAM Coloma Ballester UPF  Roc Boronat  52.023
               
 10 Tue. Dec. 6th 16:00 -18:00 HOMEWORK/HOLIDAYS         
Thu. Dec.8 th 16:00-19:00 HOMEWORK/HOLIDAYS         
               

 M2 Student Guide [here]