Introduction

SCCS 476 1

Introduction

• What is Image Processing?

• Fundamental of Image Processing

SCCS 476 2

Image & Image Processing

• Image and Pictures– An image is a single picture that represents

something

• What is image processing?– Interest in digital image processing methods stems

from two application areas: • Improvement of pictorial information for human interpretation• Processing of image data for storage, transmission, and

representation for autonomous machine perception

SCCS 476 3

Example 1: Contrast Enhancement

BEFORE AFTERGamma = 0.5

http://www.mathworks.com/access/helpdesk/help/toolbox/images/enhanc17.html

SCCS 476 4

Example 2: Sharpening

BEFORE AFTER

SCCS 476 5

Example 3: Denoising

BEFORE AFTER

http://www.mathworks.com/access/helpdesk/help/toolbox/images/enhan23b.html#14283

SCCS 476 6

Example 3: Edge Extraction

http://www.cee.hw.ac.uk/hipr/html/canny.html

SCCS 476 7

Example 4: Blurring

http://www.maths.abdn.ac.uk/~igc/tch/mx4002/notes/node99.html

SCCS 476 8

Example 5: Blurring & Sharpening

http://www.uwec.edu/walkerjs/DSP/sharpening_images.htm

SCCS 476 9

Example 6:Image Enhancement

SCCS 476 10

Example 6: Image Enhancement

SCCS 476 11

1.3 Image Acquisition and Sampling

t

A

How many points must be used to represent this curve?

At least at the rate of Nyquist rate (twice the maximum frequency in the function).

Sampling

SCCS 476 12

Undersampling• Definition: sampling signals with too few points

• Effects: aliasing (jagged edge in image)

SCCS 476 13

1.3.1 Using Light & Other Energy Sources • Light is the predominant energy source for images. • Digital images are captured using visible light, infrared,

ultraviolet, etc.

http://www.yorku.ca/eye/spectrum.gif

SCCS 476 14

1.3.2 Image Acquisition

• Camera: digital camera– CCD (Charge-Coupled Device)– CMOS (Complementary Metal Oxide Semi-

conductor)

• Flat-bed scanner– Scan row by row– Examples: Computed Axial Tomography,

MRI, etc.

SCCS 476 15

Imaging Sensors

(a) Single sensor

(b) Line sensor

(c) Array sensor

SCCS 476 16

Digital image acquisition process

SCCS 476 17

1.4 Images and Digital Images

SCCS 476 18

Image Sampling & Quantization

Quantization

Image Sampling

SCCS 476 19

Pixel

• Picture elements Pixel

SCCS 476 20

Digital Images

SCCS 476 21

Image Representation

• Consider an image as a matrix

• Intensity of pixel (x,y), f(x,y), is the member at row x and column y of the matrix

• Lexicographic ordering:– Rearrange image matrix into 1-D vector

format– Concatenate the row together

– Pixel at (x,y) is at the position x WIDTH + y

SCCS 476 22

Neighborhood

48 50 49 48 0

48 52 55 3 116

44 53 5 110 105

51 0 111 123 112

1 122 120 111 115

3 3 neighborhood

(usually is the odd number)

SCCS 476 23

1.5 Applications of Image Processing• Medicine

– Inspection and investigation of images obtained from x-rays, MRI, CAT scans

– Analysis of cell images and chromosome karyotypes• Agriculture

– Satellite/aerial views of land: determine how much land is being used

– Inspection of fruit and vegetables: distinguish good and fresh produce from old

• Industry– Automatic inspection of items on a production line– Inspection of paper samples

• Law enforcement– Fingerprint analysis

SCCS 476 24

1.6 Aspects of Image Processing (image processing algorithm)

• Image Enhancement: processing an image so that the result is more suitable for a particular application.– sharpening or deblurring– highlighting edges– improving image contrast or brightening

image– removing noise

SCCS 476 25

1.6 Aspects of Image Processing (cont)

• Image Restoration: An image may be restored by the damage done to it by known cause, for example – removing of blur caused by linear motion– removing of optical distortions– removing periodic interference

Note: (i) enhancement – make it look better,

(ii) restoration – remove damage

SCCS 476 26

1.6 Aspects of Image Processing (cont)

• Image Segmentation: Segmentation involves subdividing an image into constitute parts– finding lines, circles, or particular shapes in an

image– Identifying cars, trees buildings, or roads in an

aerial photograph

SCCS 476 27

1.7 Image Processing Task

Real-world application: A system for reading the postal codes from envelopes– Image Acquisition. First we need to produce a digital image

from a paper envelop. This can be done using either CCD camera or a scanner.

– Preprocessing. Use some image processing algorithms to obtain the resulting image more suitable for the later process. In this application it may involve enhancing the contrast, removing noises, or identifying regions likely to contain the postal code.

– Segmentation. Use some image processing algorithms to extract the region that contains postal code from the image

SCCS 476 28

1.7 Image Processing Task (cont)

– Representation and description. Extracting the particular features to differentiate between objects. Here suppose we will be looking for curves, holes, and corners that allow us to distinguish the different digits that constitute a postal code.

– Recognition and interpretation. Assigning labels to objects based on their descriptors (from the previous step) and assigning meanings to these labels. We identify particular digits, and interpret a string of digits at the end of the address as the postal code.

SCCS 476 29

1.8 Types of Digital Images

• Binary– 1 bit/pixel (black & white image)

• Grayscale – 8 bit/pixel (gray image)

• Color image:– true color: 24 bit/pixel (Red, Green Blue, 2553 colors)

• Indexed image: 8 bit/pixel (color image with 256 colors)

SCCS 476 30

Binary Image

• Two color: black and white. No gray.• Value range: 0 : black, 1 : white

http://www.mathworks.com/matlabcentral/fileexchange/loadFile.do?objectId=4052&objectType=file

SCCS 476 31

Grayscale Image

• Use black (0), white (255) and shades of gray• Value range: 0-255

SCCS 476 32

Color Image• 3 bytes for 1 pixel• R = [0,255], G = [0, 255], B = [0,255]

SCCS 476 33

Indexed Image • Mostly 1 byte for 1 pixel• Good for image composing of less than 256 colors

SCCS 476 34

Indexed Image (cont)

1 3 3 4

5 4 4 3

4 3 4 3

3 3 2 1

0.1211 0.1530 0.1234

0.1807 0.3447 0.1729

0.2627 0.2588 0.2549

0.2197 0.2432 0.2588

0.1611 0.1768 0.1924

0.2432 0.2471 0.1924

.

.

.Indices (value of the pixel)

Color Map (Palette)

Index 0Color: (0.1807, 0.3447, 0.1729)

Color: (0.1611, 0.1768, 0.1924)

SCCS 476 35

1.9 Size of Image File

• Total number of pixel = Width Height

• Size = Total number of pixel size of pixel

= Width Height #bit/pixel [bits]

= Width Height #byte/pixel [byte]

SCCS 476 36

Examples: File Size

• Binary image:– Width = 352– Height = 288– Size = ?

• Grayscale image:– Width = 352– Height = 288– Size = ?

SCCS 476 37

1.10 Image Perception

Much of image processing in concerned with making an image appear better to human beings. Therefore, we should be aware of the limitations of the human visual system.

Image perception consists of – capture the image with the eye– recognize and interpret the image with the visual

cortex in the brain

SCCS 476 38

Limitations of Human Visual System

• Observed intensities vary as to the background

SCCS 476 39

Limitations of Human Visual System (cont)

• Observation of nonexistence intensity as bars in continuously varying gray level

SCCS 476 40

Limitations of Human Visual System (cont)

SCCS 476 41

Limitations of Human Visual System (cont)

False contouring

SCCS 476 42

Limitations of Human Visual System (cont)

• Undershoot or overshoot around the boundary of regions of different intensities

– Boundary appeared brighter when seeing from dark to bright region.

– Boundary appeared darker when seeing from bright to dark region

SCCS 476 43

Limitations of Human Visual System (cont)

Mach bands