DIP_Lecture-1_2_RKJ_Introduction to Image Processing-1.pdf

dip

1. Digital Image Processing
Dr. RajibKumar Jha
Associate Professor
Depart of Electrical Engineering
Indian Institute of Technology Patna
[email protected]
1
Lecture Notes-2025

2. Lecture-1-2
Digital Image Processing: An
Introduction
2

3. Marks Distribution
•Mid Semester: 20%
•End Semester: 40%
•Project work-1: 15%
•Project work-2: 20%
•Attendance: 05%
3

4. Reference Books
•R.C.Gonzalez&R.E.Woods,“Digital
ImageProcessing”4
th
edition.
•AnilK.Jain,“FundamentalofDigital
ImageProcessing”
4

5. Contents of DIP Syllabus
•Introduction-------1
•ImageDigitization-------2
•PixelRelationships---------2
•CameraModelandImagingGeometry------1
•CameraCalibrationandStereoImaging-----2
•InterpolationandResampling------3
•ImageTransformation,FourierTransform,DiscreteCosine
Transform,K-LTransform---------5
•ImageEnhancementinspatialdomain------2
•ImageEnhancementinfrequencydomain-----2
5

6. Contents of DIP Syllabus
•ImageRestoration------3
•ImageRegistration------1
•ColourImageProcessing---3
•Image Segmentation------4
•Mathematical Morphology----3
•Image Compression------4
•Wavelet Transform & Applications---2
•Project Presentation.
6

7. Introduction
•Animageisa2-dimensionalfunctionf(x,y),wherex,yarethespatial(plane)co-
ordinatesandtheamplitudeoffatanypairofcoordinates(x,y)iscalledtheintensity
oftheimageatthatpoint.
•Ifx,yandtheamplitudevaluesoffarefiniteanddiscretequantities,wecalltheimage
adigitalimage.
•Adigitalimageiscomposedoffinitenumberofelementscalledpixelseachofwhich
hasaparticularlocationandvalue.
•DigitalImageprocessingisamethodtoperformsomeoperationsonanimage,inorder
togetanenhancedimageortoextractusefulinformationfromit.
ImageProcessinghastwoaspects:
•Improvingthevisualappearanceofimagestoahumanviewer
•PreparingImagesformeasuringdifferentfeatureslikeshape,structureetc.
7

8. Similar Areas/Doubt
•Digital Image Processing:
Input is an image-------------output is processed image
•ComputerGraphics:Herewegeneratelines,shadingsofan
objectetc.andthenputthesesasdescriptions(itisrelatedto
theimagewewanttodraw)totheinputofcomputer.Then
thatdescriptionsarecomputerprocessedinordertorepresent
thatobjectpictorially.
Input is Description-------------output is synthetic not real image
8
Comp
Comp

9. Similar Areas/Doubt
•PatternRecognition:Imageisaninputandwewanttoextractthe
meaningfuldescriptionsfromtheimage.
•Ex-Satelliteimage:howmanyclassesarethereintheimage.Thatis,
whereisthevegetationarea,waterbodyarea,forestareaetc.canbe
classified.
•Inputisanimage-----------Outputisasetofdescriptionsindicatingthe
differentclasses.
9
Comp

10. Digital Image Processing to Computer Vision
10
•ComputerVision:Afterprocessinganimageweanalysetheimageandget
moremeaningfulinformation.
•ThecontinuumfromImageProcessingtoComputerVisioncanbebrokenup
intoLow,MidandHighlevelProcesses.

11. Image Representation by Digital Computer
•AnImageisa2-Dlightintensityfunctionoff(x,y).
•Adigitalimagef(x,y)isdiscretizedbothinspatialcoordinates(sampling)and
brightness(quantization).
•Itcanbeconsideredasamatrixwhoserow,columnindicesspecifyapointinthe
imageandtheelementalvalueidentifiesgrayvalueatthatpoint.
•Theseelementsarereferredtoaspixels.
•f(x,y)=r(x,y)i(x,y); 0≤r(x,y)≤1, 0≤i(x,y)<∞
–r(x,y)reflectivityofsurfaceofcorrespondingimagepoint.
–i(x,y)intensityoftheincidencelight.
•Thelightfromthelightsourcefallsontheobjectsurface,itgetsreflected,
reachesoureyeandthenonlywecanseethatparticularobject.
•Pixelvaluesareproportionaltotheenergyradiatedbyaphysicalsource(em
waves).0≤f(x,y)<∞
11

12. Image Representation
12

13. Image Representation
13

14. Pixels

15. Pixel, Image Size, Resolution
•Everypixelkeepcolourinformation.
•Morepixelsinanimagemeanstheimagesizewillbelargebutit
doesn’talwaysmeanthatit’sabetterquality.
•Resolutionismeasuredindpiorppi(dotsperinchorpixelsperinch)
•300-PPI---Onaninchdistanceweprintout300pixelsfromtheoriginal
image.

16. Pixel, Image Size, Resolution
•Cameracaptures14.6megapixelimageswhichisaround14,600,000pixelsperimage
(14.6x1,000,000).
•Thisinformationtellsyounothingabouttheactualpixeldimensionsoftheimage–it
onlytellsyouthetotalnumberofpixelsthatcomprisetheimage.
•InmostdSLR(digitalsingle-lensreflexcamera)camera,capturedimageshavean
aspectratioof1.5.Sotheratiocomparingthenumberofpixelsalongthelongedgeto
theshortedgeis3:2.
•Eachfullsizerawimageis4672x3104pixelsindimension.So,bymultiplyingthe
numberofpixelsalongthewidthbythoseoftheheight(4672x3104=14,501,888)we
gettheactualnumberofpixelsintheimage.
16

17. Pixel, Image Size, Resolution
•Toprintanimage4x6inches
at300ppi,thenafilehasat
least4x300(1200)pixels
alongitsshortsideand6x
300(1800)pixelsonthelong
side.So,itneedstobeatleast
1200x1800pixelsinsize.
•Toprintan8x10inchesat
300ppiusethesamemath–
Theresultis2,400x3,000
pixelssizeofanimage.
17

18. Resolution
For commercial printer 300ppi is standard

19. Low and High Resolution Image

20. Resolution

21. Resolution/Image Size
•All theimage resolution formats reach up to
1280x720, they are high definition (HD).
•CurrentHDresolutionformatsinclude1.0
megapixel(720p),1.3megapixel(960p),2
megapixel(1080p),3megapixel,5megapixel,
8megapixel(4KUHD),12megapixel,33
megapixel(8KUHD).

22. HD (high degree of detail) Resolution

23. 23

24. Electromagnetic Spectrum
24

25. 25

26. 26
•It is used for treatment of cancerous cell in our body without the use of the surgery.
•It is used in industries to kill the harmful bacteria, organism like yeast etc.
•Like x –rays, it is also used to disinfect medical instruments.
•It is used to detect brain and heart abnormalities.
•Gamma rays are used by Engineers, since they can penetrate better than X-rays, to
look for cracks in pipes and aircraft parts.
•Gammaraysaremostlyusedin
theradiotherapy/radio-
oncologytotreatcancer.
•Gammarayscankilllivingcells
anddamagemalignanttumor.

27. 27
UsesGammarayscankilllivingcells,theyareusedtokillcancercellswithouthavingtoresorttosurgery.
Thisprocedureiscalled"Radiotherapy",

28. 28

29. X-ray Images
29

30. X-ray image
30

31. X-ray image
31
Tomosynthesisor “3D” mammography is a new type of digital x-ray
mammogram which creates 2D and 3D-like pictures of the mammogram

32. Ultraviolet-ray
•Ultravioletphotographyisaphotographicprocessofrecording
imagesbyusinglightfromtheultraviolet(UV)spectrumonly.
•Diagnosticmedialimagesmaybeusedtodetectcertainskin
disordersorasanevidenceofinjury.
32

33. UV Ray Image
•Ultraviolet(UV)lightisacommonlyusedmethodforkillingbacteriaand
•Veryhighlevelsofultravioletlightdisinfectsareascontaminatedwith
COVID-19etc.
33

34. Ultraviolet photography
34

35. Ultraviolet-ray
35
Thebiggestcontributortoagingskinisthesunexposureduringour
lifetimes.

36. Visible Light Spectrum
36

37. Visible ray
37

38. Infrared images
38

39. Normal Image/Infrared Images
39

40. Infrared images
40

41. Night Vision
41

42. IR Image
42

43. Why do we need Image Processing ?
•It is motivated by some major applications:
–Improvement of pictorial information
–autonomous machine application
–Efficient Storage and transmission
–Object detection & recognition
43

44. Why Image Processing
•Google,
•Amazon,
•Facebook,
•Twitter,
•Cellphone company,
•Computer Vision,
•Machine Learning,
•Deep learning
44

45. Why do we need to process the images ?
•Improvementofpictorialinformationforhumanperception
–Imagecontentenhancement:whateverimageweget,wewant
toenhancethequalityofanimagesothattheimagewillhave
abetterlook.Somemethodsare
•Noisefiltering&ImageEnhancement
•Debluringofblurredimage
•ClassifyingpixelsofRemotesensingimages
45

46. Noise Filtered image
46

47. Image Enhancement
47

48. OutputImage
Image Enhancement
48
Input Image

49. Input Image
Image Enhancement
49
OutputImage

50. Image Enhancement
50
Input Image Output Image
Input Image Output Image

51. 51

52. 52

53. Image Deblurring
53

54. Blurred and De-blurred Image
54

55. Blurred and De-blurred Image
55

56. Classifying pixels of Remote sensing images
56

57. Remote sensing image
57
•Study that whether the river has changed its path,
•Study what is the growth of vegetable over a certain region etc.

58. Remote Sensing image to Count Trees
58
•Build a city over certain region with help of satellite image.
•One can determine that where the residential area, industrial area etc. have
to be grown.

59. •Imageprocessingforautonomousmachineapplication
–Industry Machine vision for product assembly and inspection
–Automated target detection and tracking
–Finger print recognition
–Medical Applications
59

60. Automated Inspection
60

61. Fingerprint Enhancement
61

62. Medical Imaging
62

63. Medical Imaging
•For different types of fatty images, cancerous
regions look like (a) to (c)
63

64. Medical Imaging
64

65. •Video(displayingofimagesequence)SequenceProcessing:
–Detectionandtrackingofmovingtargetforsecuritysurveillance.
–Monitoringthemovementsoforganboundariesinmedical
applicationsetc.
65

66. Movement detection
66

67. Movement detection
67

68. •Efficientstorageandtransmission:
–Animageusuallycontainslotofredundancythatcanbeexploitedtoachieve
compression.
–Ifimagedataiscompresseswithoutinformationlossthenitisveryusefulfordata
compression&transmission.
–Redundancy
•Pixelredundancy
•Codingredundancy
•Psychovisualredundancy
68
•ImagekeepsInformationandRedundancyboth.
•Soprocessonimage,removetheredundancyandkeeptheinformation.

69. 69
•Psychovisualredundancyarisesduetotheproblemofperception.
•Oureyesaremoreresponsivetoslowandgradualchangesofillumination
thanperceivingfinerdetailsandrapidchangesofintensities.
•Hence,towhatextentweshouldpreservethedetailsforperceptionor
compromiseonthequalityofreconstructedimageisessentiallycarried
outbyexploitingthepsychovisualredundancy.
Psychovisualredundancy

70. Image Compression
70

71. 71
Image Compression

72. Fundamental steps of digital image processing
72

73. Fundamental steps of digital image processing
•Imageacquisition:Itcouldbeassimpleasbeinggivenanimagethatis
alreadyindigitalform;
•Imageenhancement:Ideaistobringoutdetailsthatareobscured,or
simplytohighlightcertainfeaturesofinterestinanimage;
•Imagerestoration:thatdealswithimprovingtheappearanceofan
image;
•Colorimageprocessing:includecolormodelingandprocessingina
digitaldomain,etc;
•Waveletsandmultiresolutionprocessing:Thesearethefoundationfor
representingimagesinvariousdegreesofresolution;
•Compression:Reducingthestoragerequiredtosaveortransmitan
image;
73

74. Fundamental steps of digital image processing
•Morphologicalprocessing:Extractimagecomponentsusefulinthe
representationanddescriptionofshape;
•Segmentation:Dividesanimageintoitsconstituentpartsorobjects;
•Representationanddescription:Transformsrawdataintoaform
suitableforsubsequentcomputerprocessing;
•Objectrecognition:Itistheprocessthatassignsalabelintoobject
basedonitsdescriptor;
•Knowledgebase:Itindicatesregionsofanimagewheretheinformation
ofinterestisknowntobelocated
74

75. Thanks
75