CSC 381/481 Quarter: Fall

1
CSC 381/481 Quarter Fall 03/04

• Daniela Stan Raicu
• Email draicu_at_cs.depaul.edu
• Homepage http//facweb.cs.depaul.edu/dstan
• School of CTI, DePaul University

2
Outline

• Class overview
• Introduction to Image Processing
• Main Concepts
• Applications
• Getting started with Matlab

3
Class Overview

• Syllabus
• Summary of the course
• Textbook
• Prerequisites
• Grading
• Software
• Homework/Programming Assignments and Final
  Project
• School Policies
• Office hours
• Tuesday, 445pm-545pm
• Thursday, 330pm-400pm
• Office location CTI 471, Loop campus

4
What is a Digital Image?
A digital image is a computer representation of
a photograph. It consists of a two-dimensional
array of integer numbers. The numbers are always
positive because they represent the variations in
the brightness (reflected light energy) in the
corresponding photograph. Each element of the
two-dimensional array representing a digital
image is called pixel a short form for picture
elements.
a(1,4)110 (the pixel at coordinates row1
and column4 has the integer value110). The
process of representing the value of the pixel
is called Quantization.
Sampling/ Digitization
a(x,y), xof rows, yof columns
5
What is a Digital Image? (cont.)
Low-level interpretation a digital image is an
array of pixel values with no meaning. High-level
interpretation an image is worth a thousand of
words. The challenge of extracting meaningful
patterns from such data sets (array of pixels)
lead to the research and development in the area
of multimedia data mining.

Ocean Scene
Sky
Mountains
Grass
6
What is Digital Image Processing?
Digital image processing is nothing but
computer-based manipulation of digital images. It
can be considered as an analog of photographic
darkroom. The input to a digital image processor
is an image. The output of a digital image
processor is also an image. This distinction is
important to remember because many computer
systems deal with digital images with different
goals.
Output edge blood image
Input Blood Image
Edge detection demo under MATLAB gtgtdemo
7
What is Digital Image Analysis?
Digital image analysis deals with the
computation of image properties such as the
number of objects, for example the blood cells in
a cell image, their image locations, object
areas, and other desired properties. Thus, the
input to a digital image analyzer consists of an
image and the output is comprised of a set of
numbers representing certain image properties.
Output labeled steel grain image
Input Steel grain image
Region Labeling of steel grains under MATLAB
demos gtgtdemo
8
What is Computer Vision?
Computer vision / machine vision also deals with
images. The goal of a computer vision system is
to extract geometrical information from images
for actions such as locomotion and object
manipulation. A computer vision system must have
certain digital image processing and analysis
capabilities in achieving its goal. Additionally,
a vision system has a large knowledge-base
pertaining to its domain of operation. The
knowledge-base allows the vision system to
evaluate various action options and plan the
future actions.
Tracking people for monitoring shopping
Precincts Video images ---Fig 1. Edge
detection---Fig 2 Then, using a model of the
major body parts, anyone in the picture can be
located. Once a person is located, they can be
tracked by predicting and testing the possible
movements from a range of legal articulations of
the model ---Figs 3 and 4. Then any deviation
from the usual paths can raise an alarm.
9
What is Computer Graphics?
The output of a computer graphic system is also
an image. However, a graphic system produces an
output image based the geometrical and surface
descriptions of a set of objects, lighting
sources, and the geometry of an imaginary camera.
Computer graphics / image synthesis shares
several concepts and techniques with digital
image processing and analysis.
10
What is Visual Computing?

• Visual computing is an emerging field of study
  arising due to the confluence of
• computer graphics
• Creating artificial images and videos from
  scratch
• digital image processing and analysis
• Analysis of images and videos of the real world
• computer vision
• Therefore, it deals with synthesis, processing,
  transformation, and indexing of digital images
  and videos.

11
Why is Visual Computing hot?

• Applications Market
• Medical imaging
• The global market is worth in excess of 1
  billion and is growing at about seven per cent
  per year.
• Example Planning Radiation Therapy

When planning radiation therapy of brain tumors
from MR scans of the head, it is necessary to
apply to most effective dose of radiation to the
tumor, yet cause least damage to the surrounding
tissue.
One approach is for the clinician to select a
point within an area of interest, such as the
tumour. The system then locates the surrounding
points which possess similar characteristics,
within a given range of variability. This is
repeated for other regions, and other images, to
produce a 3D model. Given this model, an optimal
treatment plan can be generated.
12
Why is Visual Computing hot? (cont.)

• Applications Market
• Industrial Applications
• Industrial inspection
• Fruit grading
• Inspecting machined parts
• Remote Sensing and Space Applications
• Land management
• Crop classification
• Surveying by satellite
• Military Applications
• Recognition of enemy aircraft from radar screens
• identification of targets from satellite
  photographs

13
More on visual computing
Seeing is Believing Computer Vision and
Artificial Intelligence, by Christopher O.
Jaynes, http//www.acm.org/crossroads/xrds3-1/visi
on.html Good places to start reading about
computer vision and artificial intelligence http
//www.aaai.org/AITopics/html/vision.html Applica
tions of computer vision and careers in computer
vision http//www.bmva.ac.uk/apps/index.html Glo
ssary http//www.imagelabs.com/gl.php
14
Getting started with Matlab

• Matlab
• Matrix algebra laboratory
• A technical computing environment for high
  performance numeric computation and
  visualization.
• Matlab also features a family of
  application-specific solutions called toolboxes.
• Toolboxes are comprehensive collections of
  Matlab functions (M-files) that extend its
  environment in order to solve particular classes
  of problems.
• Example of toolboxes
• - Image Processing
• - Signal Processing
• - Neural Networks
• - Statistics, etc

15
Getting started with Matlab

• Matlab Usage
• In order to start Matlab, simply type
• matlab
• If you are using a workstation with X
  capabilities, just start matlab from a console or
  shell of the workstation you are using Matlab
  will notice that you are on a workstation and
  should open a graphics window.
• MATLAB and the Image Processing Toolbox are
  installed in the CTI Software Development (SD)
  lab (on the sixth floor of the CTI building)
• Matlab 6.5 Student version release 13
• http//www.mathworks.com/products/education/studen
  t_version/sc/index.shtml

16
Getting started with Matlab

• Matlab Documentation
• built-in help
• Matlabs help command
• to access its built-in help
• Matlabs lookfor command
• to access its built-in help
• hardcopy manuals
• on-line Matlab reference
• documentation using a WWW
• document viewer this is fully
• indexed, contains color graphics,
• and includes hypertext links between
• related topics.
• Matlabs doc to start the www document viewer

17
Getting started with Matlab

• Matlab Documentation (cont.)
• Image processing toolbox documentation
  http//www.mathworks.com/access/helpdesk/help/tool
  box/images/images.shtml
• Matlab Primer
• The purpose of this Primer is to help you begin
  to use Matlab as you read the Primer, freely
  experiment with examples in order to get hands-on
  experience.
• http//math.ucsd.edu/driver/21d-s99/matlab-prime
  r.html
• Tutorial code snippets
• http//www-cse.ucsd.edu/sjb/classes/matlab/matla
  b.intro.html
• Matlab FAQ http//www.mit.edu/pwb/cssm/

18
Getting started with Matlab
Matlab Demos demos
19
Getting started with Matlab
Image Processing Matlab Demos