Department of Computer Engineering, CMU Chapter 1 ... · Department of Computer Engineering, CMU 3 5 Overview Early days of computing, data was numerical. Later, textual data became

Department of Computer Engineering, CMU

1

Chapter 1: Introduction

Lecturer: Wanasanan ThongsongkritEmail : [email protected]

Office room : 410

2

Text book & Reference bookTextbook:

Rafael C. Gonzalez and Richard E. Woods, "Digital Image Processing, 2nd edition", Prentice Hall, 2001.

Reference book:Rafael C. Gonzalez and Richard E. Woods, "Digital Image Processing", Addison Wesley 1993.Milan Sonka, Vaclav Hlavac and Roger Boyle, "Image Processing Analysis, and Machine Vision 2nd edition", PWS Publishing, 1998.


2

3

Course ContentChapter 1: IntroductionChapter 2: Digital Image FundamentalsChapter 3: Image Enhancement in the Spatial DomainChapter 4: Image Enhancement in the Frequency DomainChapter 9: Morphological Image ProcessingChapter 10: Image SegmentationChapter 11: Representation and DescriptionChapter 12: Object Recognition (introduction)Chapter 8: Image Compression (introduction)

This class is limited on gray level 2-D images.

4

Grading System

Midterm examination 35% (first 4 chapter

Final examination 40% (the rest)

Assignment/Quiz/Report 25% Warnings:

A quiz may be given without being informed before.Copying assignment is prohibited. Delay of submission influences on marks


3

5

OverviewEarly days of computing, data was numerical.Later, textual data became more common.Today, many other forms of data: voice, music, speech, images, computer graphics, etc.Each of these types of data are signals. Loosely defined, a signal is a function that conveys information.

6

Relationship of Signal Processing to other fields

As long as people have tried to send or receive through electronic media : telegraphs, telephones, television, radar, etc. there has been the realization that these signals may be affected by the system used to acquire, transmit, or process them.Sometimes, these systems are imperfect and introduce noise, distortion, or other artifacts.


4

Understanding the effects these systems have and finding ways to correct them is the fundamental of signal processing.Sometimes, these signals are specific messages that we create and send to someone else (e.g., telegraph, telephone, television, digital networking, etc.).That is, we specifically introduce the information content into the signal and hope to extract it out later.

Sometimes, these man-made signals are encoding of natural phenomena (audio signal, acquired image, etc.), but sometimes we can create them from scratch (speech generation, computer generated music, computer graphics).Finally, we can sometimes merge these technologies together by acquiring a natural signal, processing it, and then transmitting it in some fashion.


5

Transmitted codes of image

Decoded Decompressed

Displayed to create another signal (visible

light of the display)

received by eyes

Interpreted in some fashion by our brain

Acquire natural image

Enhance the picture Compress for transmissionEncode and

transmit over digital network

Recipient:

Sender:

From acquisition to interpretation, the initial signal may be transformed, modified, and retransmitted numerous times.

10

Concerned fields:Digital CommunicationCompressionSpeech Synthesis and RecognitionComputer GraphicsImage ProcessingComputer Vision


6

11

What is Image Processing?Image processing is a subclass of signal processing concerned specifically with pictures.Improve image quality for human perception and/or computer interpretation.

Image Image Processing Better Image

12

Several fields deal with imagesComputer Graphics : the creation of images.Image Processing : the enhancement or other manipulation of the image – the result of which is usually another images.Computer Vision: the analysis of image content.


7

13

AIComputer GraphicsDescription

Computer Vision

Image ProcessingImage

DescriptionImageInput/Output

Computer Vision, Image Processing and Computer Graphics often work together to produce amazing results.

Several fields deal with images

14

2 Principal application areasImprovement of pictorial information for human interpretationProcessing of image data for storage, transmission, and representation for autonomous machine perception


8

15

Ex. of fields that use DIPCategorize by image sources

Radiation from the Electromagnetic spectrumAcousticUltrasonicElectronic (in the form of electron beams used in electron microscopy)Computer (synthetic images used for modeling and visualization)

16

Radiation from EM spectrum

EM waves = a stream of massless (proton) particles, each traveling in a wavelike pattern and moving at the speed of light.Spectral bands are grouped by energy per photon

Gamma rays, X-rays, Ultraviolet, Visible, Infrared, Microwaves, Radio waves


9

17

Gamma-Ray ImagingNuclear Image

(a) Bone scan(b) PET (Positron emission tomography) image

Astronomical Observations.

(c) Cygnus LoopNuclear Reaction

(d) Gamma radiation from a reactor valve

dcba

18

X-ray ImagingMedical diagnostics

(a) chest X-ray (familiar)(b) aortic angiogram(c) head CT

Industrial imaging(d) Circuit board

Astronomy(e) Cygnus Loop

ecb

da


10

19

Imaging in Ultraviolet BandLithographyIndustrial inspectionMicroscopy (fluorescence)

(a) Normal corn(b) Smut corn

LasersBiological imagingAstronomical observations

(c) Cygnus Loop

cba

20

Imaging in Visible and Infrared Bands

AstronomyLight microscopy

pharmaceuticals(a). taxol (anticancer agent)(b). cholesterol

Microinspection to materials characterization

(c). Microprocessor(d). Nickel oxide thin film(e). Surface of audio CD(f). Organic superconductor

fc

edba


11

Remote sensing Thematic bands in visual and infrared regions

To monitoring environmental conditions on the planet

NASA’s LANDSAT: Washington DC

Remote sensing : Weather observation and prediction

Multispectral image of Hurricane Andrew from satellites using sensors in the visible and infrared bands


12

Remote sensing : Nighttime Lights of the World(provides a global inventory of human settlements)

Infrared satellite images of the Americas.

Infrared satellite images of the remaining populated part

of the world

Industry : visual spectrum(automated visual inspection of manufactured goods) e

fcb

da

(a). A circuit board: inspect them for missing parts(b). Pill container: look for missing pills(c). Bottles : look for bottles that are not filled up to an acceptable level(d). Bubbles in clear-plastic product : detect unacceptable number of air pockets(e). Cereal : inspection for color and the presence of anomalies such as burned flake.(f). Image of replacement lens for human eye : inspection of damaged or incorrectly manufactured implants


13

Law enforcement : visual spectrum

(a). Thumb print: automated search of a database for a potential matches(b). Paper currency : automated counting / reading of the serial number for tracking and identifying bills(c) and (d) Automated license plate reading

cba

26

Imaging in Microwave BandImaging radar : the only way to explore inaccessible regions of the Earth’s surfaceRadar image of mountains in southeast TibetNote the clarity and detail of the image, unencumbered by clouds or other atmospheric conditions that normally interfere with images in the visual band.


14

27

Imaging in Radio BandMedicine

Magnetic resonance image (MRI) : 2D picture of a section of the patient (any plane)

(a) knee(b) spine

Astronomy

28

Acoustic ImagingGeological applications : use sound in the low end of the sound spectrum (hundred of Hz)

Mineral and oil exploration

Cross-sectional image of a seismic model. The arrow points to a hydrocarbon (oil and/or gas) trap

(bright spots)


15

29

Ultrasound ImagingManufacturingMedicine

(a) Baby(b) Another view of baby(c) Thyroids(d) Muscle layers showing lesion

dcba

30

Generated images by computer

Fractals : an iterative reproduction of a basic pattern according to some mathematical rules

(a) and (b)3-D computer modeling

(c) and (d)

dcba


16

31

3 types of computerized process

Low-level : input, output are imagesPrimitive operations such as image preprocessing to reduce noise, contrast enhancement, and image sharpening

Mid-level : inputs may be images, outputs are attributes extracted from those images

SegmentationDescription of objectsClassification of individual objects

High-level :Image analysis

32

Fundamental steps

Problemdomain

Chapter 2

ImageAcquisition

Chapter 2

ImageAcquisition

Chapter 3&4

ImageEnhancement

Chapter 3&4

ImageEnhancement

Chapter 5

ImageRestoration

Chapter 5

ImageRestoration

Chapter 6

Color ImageProcessing

Chapter 6

Color ImageProcessing

Chapter 7Wavelet andMultiresolutionProcessing

Chapter 7Wavelet andMultiresolutionProcessing

Chapter 8

Compression

Chapter 8

Compression

Chapter 9

Morphological

Processing

Chapter 9

Morphological

Processing

Chapter 10

Segmentation

Chapter 10

Segmentation

Chapter 11

Representation&Segmentation

Chapter 11

Representation&Segmentation

Chapter 12

ObjectRecognition

Chapter 12

ObjectRecognition

Knowledge base

Output of these processes generally are images

Ou

tpu

t of

th

ese

proc

esse

s ge

ner

ally

are

imag

e at

trib

ute

s


17

33

Image Acquisition:

An image is captured by a sensor (such as a monochrome or color TV camera) and digitized.If the output of the camera or sensor is not already in digital form, an analog-to-digital converter digitizes it.

34

CameraCamera consists of 2 parts

A lens that collects the appropriate type of radiation emitted from the object of interest and that forms an image of the real object a semiconductor device –so called charged coupled device or CCD which converts the irradiance at the image plan into an electrical signal.


18

35

Frame GrabberFrame grabber only needs circuits to digitize the electrical signal from the imaging sensor to store the image in the memory (RAM) of the computer.

36

Image EnhancementTo bring out detail is obscured, or simply to highlight certain features of interest in an image.

Example:


19

37

Image RestorationImproving the appearance of an imageTend to be based on mathematical or probabilistic models of image degradation

Example:

Restored imageDistorted image

38

Color Image ProcessingGaining in importance because of the significant increase in the use of digital images over the InternetHowever, our lecture is limited to gray-level image processing


20

39

WaveletsFoundation for representing images in various degrees of resolution.Used in image data compression and pyramidal representation (images are subdivided successively into smaller regions)

40

CompressionReducing the storage required to save an image or the bandwidth required to transmit it.Ex. JPEG (Joint Photographic Experts Group) image compression standard.


21

41

Morphological processingTools for extracting image components that are useful in the representation and description of shape.

42

Image Segmentation

computer tries to separate objects separate objects from the image backgroundfrom the image background.It is one of the most difficult tasks in DIP.A rugged segmentation procedure brings the process a long way toward successful solution of an image problem.Output of the segmentation stage is raw pixel data, constituting either the boundary of a region or all the points in the region itself.


22

43

Representation & DescriptionRepresentation make a decision whether the data should be represented as a boundary or as a complete region.

Boundary representation focus on external shape characteristics, such as corners and inflections.Region representation focus on internal properties, such as texture or skeleton shape.

44

Representation & Description1 connected component, 1 hole

1 connected component, 2 holes

Representation + Description

transform raw data

a form suitable for the Recognition

processing


23

45

Recognition & InterpretationRecognition the process that assigns a label to an object based on the information provided by its descriptors.Interpretation assigning meaning to an ensemble of recognized objects.

46

Knowledge basea problem domain detailing regions of an image where the information of interest is known to be located. Help to limit the search


24

47

Not all the processes are needed. Ex. Postal Code Problem

Desired output = alphanumeric characters

Segmentation Representationand description

Knowledge

base

Recognition

and

Interpretation

Problemdomain

Result

Pieces of mail

Image acquisition

Preprocessing

Documents

Department of Computer Engineering, CMU Chapter 1 ... · Department of Computer Engineering, CMU 3 5 Overview Early days of computing, data was numerical. Later, textual data became