CS 696: Applied Computer Vision

Spring 2016, MW 2:00-3:15PM, AH-2112

Instructor: Xiaobai Liu

Updates on 1/12/2016

Course Description

This course shall cover concepts, algorithms, and practices of computer vision that are widely used for solving challenging problems. The topics include image processing, feature detection/matching, segmentation, image alignment and stitching, structure from motion, recognition, and tracking.

PreRequisites

This course requires programming experience as well as linear algebra, basic calculus, and basic probability. Previous knowledge of visual computing will be helpful.

The following courses (or equivalent courses at other institutions) are helpful prerequisites:
CS 310 Data Structures
CS 559 Computer Vision
CS 596 Machine Learning

Some of the course topics overlap with these related courses, but none of the assignments will.

Reading Materials

Textbook: "Computer Vision: Algorithms and Applications" by Richard Szeliski. The book is available for free online or available for purchase.

Others: papers or articles specified in class meetings.

Grading

  • 40% One semester-long project (SP), two submissions
    • 20% mid-term submission
    • 20% final-term submission
  • 55% Six homework assignments (HA), each HA is mixture of programming and quizs.
    • 5% HA1
    • 10% HA2
    • 10% HA3
    • 10% HA4
    • 10% HA5
    • 10% HA6
  • 5% attendance (signature required)

Semester project(SP) will be announced in the first two weeks. Students team with each other to complete two submissions for mid-term and final-term, respectively. Each team consists of no more than 3 students. Topics are selected by students with the assistance of the professor.

Turn in all homework assignments by email before11:59pm on the due date. No hardcopy is required. One will lose 10% each day for late submission. However, there will be only three 'late day' for the whole semester.

Important Date

Semester Project (SP) for Mid-term, due at 11:59pm, March 24, 2016

Semester Project (SP) for Final-term, due at 11:59pm, May 1, 2016

Programming Language

It is strongly recommended that all projects be completed in Matlab. All exemplar codes will be provided for Matlab . Students may try to implement the same codes in other languages which is in general way more difficult and time-consuming.

This course however doesn't cover the programming skills in matlab. Students may find a good tutorial in this link: Matlab Tutorial

Office Hours and Contact information

Email is the best way to communicate: Xiaobai.liu@mail.sdsu.edu.

Office Hours: 10:50-11:50 am M/W, GMCS 542, apointment by email;

Tentative Syllabus

Syllabus and slides will be available in SDSU blackboard after class meetings.

HA: Homework Assignment;

SP: Semester Project.

Class Topic Slides Reading Assignments
1 Computer vision and its applications   Szeliski 1
Image Formation and Filtering
2 Geometry and camera models   Szeliski 2.1, especially 2.1.5 HA1 Out
  Color space   Szeliski 2.2 and 2.3 SP Out
3 Image filtering   Szeliski 3.2  
  Frequency Domain   Szeliski 3.4 HA2 Out
4 Image Pyramids   Szeliski 3.5.2 and 8.1.1
Feature Detection and Matching
  Edge detection   Szeliski 4.2
5 Interest points and corners   Szeliski 4.1.1  
  Local image features   Szeliski 4.1.2 HA3 Out
6 Feature matching and hough transform   Szeliski 4.1.3 and 4.3.2
  Model fitting and RANSAC   Szeliski 6.1
Multiple Views and Motion
7 Stereo System   Szeliski 11
  Epipolar Geometry and Structure from Motion   Szeliski 7 HA4 Out
8 Feature Tracking and Optical Flow   Szeliski 8.1 and 8.4  
Machine Learning Tool: clustering and classification
  Machine learning intro and clustering   Szeliski 5.3
9 Machine learning: clustering continued   Szeliski 5.3  
  Machine learning: classification   HA5 Out
Recognition
10 Recognition overview and bag of features   Szeliski 14
  Large-scale instance recognition   Szeliski 14.3.2
11 Detection with sliding windows: Viola Jones   Szeliski 14.1, 14.2  
  Class Senimar: Modern object detection   Szeliski 14.1,
12 Class Senimar: Modern boundary detection and Pb   Szeliski 4.2  
Segmentation
  Region grouping with k-means   HA6 Out
13 Mean-shift for region grouping      
  Grpah based Methods      
Tracking
14 CamShift tracker  
  KLT tracker  
 
15 Class Senimar: recent advances in computer vision  
  Class Senimar: student presentations for final-term  

Links

·         CV Online

·        OpenCV (open source computer vision library)

·        Weka (Java data mining software)

·        Compiled list of image datasets

·        Object recognition databases (list compiled by Kevin Murphy)

·        Various useful databases and image sources (list compiled by Alyosha Efros)

·        Netlab (matlab toolbox for data analysis techniques, written by Ian Nabney and Christopher Bishop)

·        Oxford Visual Geometry Group (contains links to data sets and feature extraction software)

·        Computer vision conferences

·        Annotated computer vision bibliography

·        Face recognition homepage

·        Computer vision research groups

·        Vision related links on AAAI.org page

·        Linear algebra review / primer by Martial Hebert