/cvclasses16

Computer Vision Classes 2016

Primary LanguageC++

Computer Vision Classes 2016

Overview

This is a Computer Vision classes prepared for the 2nd-year M.D. students. Course includes topics covering basic areas of computer vision. It takes 8 seminars-discussion lessons and 8 practical lessons. Most practical lessons will be based on using of the OpenCV v3. library. All materials and tasks will be hosted in this repository and available for all students. All program code is written under Visual Studio 2015 with v120 toolset.

First Steps

Please ensure you have following instruments and settings to start the work:

  • Visual Studio 2013 (2015 with v120 Platform Toolset)
  • Git
  • OpenCV v3.+
  • Image Watch plugin for VS
  • Camera to execute demo applications
  • Setup Environment variable: OPENCV_DIR = C:...\opencv\build\
  • Put opencv_world300.dll into Release folder after first build and opencv_world300d.dll into Debug folder.

Required Knowledge

  • Basics of Digital Signal Processing
  • Basics of Digital Image Processing
  • C++ experience
  • Git experience

Feedback Form

Please feel free to leave your feedback about lecture, home task or course overall via anonymous feedback form (in russian).

Topics

1. Introduction into CV

Organizational questions, course plan and basics of the OpenCV library.

2. Image Segmentation

Overview and discussion of set of widely-used image segmentation approaches and algorithms, how to test and verify them and public testing images databases.

3. Edge Detection

Overview and discussion of set of edge detection algorithms based on 1st, 2nd derivative and Canny. How to test and verify them and public testing images databases.

4. Movement Detection

Overview and discussion of set of basic background subtraction non-recoursive (Mean/Median Filter) and recoursive (MinMax, 1G, GMM, etc.) algorithms. How to test and verify them and public testing images databases.

5. Feature Detection

Overview and discussion feature definition and properties. Corner Detection algorithms by Moravec, Haris, FAST, etc are described as wall as LoG, DoG blob detectors.

6. Feature Matching

Overview and discussion of SIFT and SURF detectors/descriptors. Basic image transformation models (Affine, Perspective) and feature matching approaches is described.

7. Feature Tracking

Overview and discussion Feature Tracking algorithms based on Lucas-Kanade ideas published in 1981.

8. Stereovision

TODO

References

  1. http://opencv.org/
  2. TODO3