1 CS4243 Computer Vision & Pattern Recognition A/Prof NG Teck Khim

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 29

CS4243

Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 1


• Lecture Monday 6:30pm – 8:30pm

• Lab 20%

• Course Project 40%


(4 persons to a team, TBC)

• Final Exam 40%

CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 2


venue: AS6 04-21

• Tuesday 2pm – 4pm

• Tuesday 4pm – 6pm

• Tuesday 6:30pm – 8:30pm

• Thursday 4pm – 6pm

• Friday 10am – 12pm


CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 3
• Open book

• 29 Nov 2017 6pm

CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 4


Lecturer:
A/Prof Ng Teck Khim
office: COM2 03-28
ngtk@comp.nus.edu.sg

Teaching Assistants:

Hu Sixing Lin Jiahao


hu.sixing@u.nus.edu jiahao.lin@u.nus.edu

CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 5


Time: Every Monday 4pm – 5pm

Venue: COM2-03-28

CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 6


• Understand the concepts and see the intuitions

• Mathematics are needed to understand the


concept.
• This is a course on computer vision, not mathematics.
• But you need to know some basic skills in mathematics
to understand the concept.
• Heavy use of Linear Algebra and some Calculus.
• You will not be tested on complicated mathematical
manipulations.

CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 7


Overview of Course

CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 8


“computer vision ≠ image processing”

Computer Vision: computer Image Processing:


Image analysis and vision study of images from
study of image signal processing
formation geometry image viewpoint
eg. 3D modeling and processing eg. image compression,
scene understanding image enhancement,
etc. image restoration

pattern
recognition

CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 9


Some image processing examples …

note: some of the pictures in this set of slides are taken from internet

CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 10


Before After

CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 11


Histogram Equalization

CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 12


CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 13
CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 14
Some computer vision examples …

CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 15


Feature Tracking

CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 16


Image patch matching
(feature correspondence)
Eg. SIFT, SURF, etc.

Difficulties:
• Flag is non-rigid
• Bottle is specular
• Bottle and gray giraffe
are feature-less
• Background noise
CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 17
Camera Calibration

CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 18


Camera Projection Models

3D Scene point
Camera coordinate kf sp - t f
system
if
tf jf sp
Z t f : camera position w.r.t. world
World coordinate system
coordinate X
s p : 3D position of scene point
system
w.r.t world coordinate system
Y
s p - t f : 3D position of scene point
w.r.t. camera coordinate system
CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 19
3D 2D
graphics

vision

CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 20


CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 21
CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 22
CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 23
http://www.pokemon.com/us/pokemon-
video-games/pokemon-go/

http://www.pocket-lint.com/news/135688-what-is-magic-
leap-and-why-might-it-kill-all-screens

Magic Leap
CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 24
IS3261 A/Prof NG Teck Khim 25
IS3261 A/Prof NG Teck Khim 26
Course Project (2 Years Ago)
https://drive.google.com/file/d/0B9WyrfZ8p4DuZUc4WGx5YnZSdWc/view

CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 27


Course Project (last year)
https://www.youtube.com/watch?v=sYsTVgOQUe4&app=desktop

CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 28


Summary of Course Overview

• We will do image processing, computer vision and a


bit of pattern recognition.

• Nature of materials in this course:


• Theory based (hard truth J)
• Heuristics (experience)

• Need mathematics
• Linear algebra (a lot)
• Calculus (a little bit)

CS4243 Computer Vision & Pattern Recognition A/Prof Ng Teck Khim 29

You might also like