To SIGITAL IMAGE F\ Syypeo 4 UNDAMENTALS § IMAGE TRANSFORMS ° ® Image: Image is a two dimensional function FO where LOW) o Wy Tmage: Tog 4 © @? Me spatial coordinates Example : Here (%0s40) (1141) ave called spatial coordinate ¥ > io digital image Natty Yes rimages are very powerfel tool in communication” LL a I? image may be,, white & black (on colour "peasy? 2 HS amplitode af fang point f(y) aives intensity (= = wy qay level at that pant : ' S At different pants we hove different intensity of 4 h > gray level valves -of an image = | 2 eA simple image model R 9 * Image must suitable fox computer processing for this an J? image Ply) most be dig talized both in spatial and armplitede. * 2 * Digitalized of spatial cocrchnates (uy) 1s called sampling (ov) aa image sampling. a # Digitalization oF arnplitude value is called quantization ov gray 34 7 level quantization. a method to convert an image into Lge) # a Trrage processing: Tt 15 3% digital fom and pextosm some operations on it in otdey te B get an enhanced image. ibis -a process in which input ish | 1g > and olp may be image oF characters associated with thatinge @ Tmage processing steps tt follows thiee steps > Step 42 Importing “the image Via ‘image acquistion ‘tool > stepr: analyzing & manuplating -the image a 1 > step a:output is the last stage in uhich yesult can be alters 7 > image. = 2 w Digital irnage a « When all values of (414) & amplitude of £ ave finite g ~ = 2 discrete values then it is called as cligital image g * Intensity at each point ‘of image is finite & discvete “ 5 Valve.® Digital image processing * It veters to the processing of digi . v computers. pa 4 digital image to digital # Digital image consists of Finite no. oF elements with particolay & i locations and values - This elements axe commonly called a5 image € elements or pitel elements. afferent VOLES qm wéach pixel is represented by K bits a pinel can have 2 di be r Fu. @ bit image 28= 256 (0-255) | i. limited # Human vision ig the most advanced sensor, but Wis eengor vr 40 vision band of elect tomagnetic spect vom #To Overcome this disadvantage . image machines are Used 2 This machine covers all the bands in the EM spectum. cat je Hence we con say that a dip covers wide & varity fields oF OPPICNONS due 10 abue reasons #image + Any visval object that's modified of altered by a computes oy an imagioary object created using a computey x picture +A drawing, parnting . oF avteoork AVE 15 also describes anything created using @ camera # Advantages of OrP created on a computer.A pict or scanner color image —_ wenelets of 5, anage morphological ienage processing oa [0m “compression puocessing Teege restoration Tonage enhanc errent mage seqmentation € © e e e t e ce t ® t * zrttrccr- knowledge base Representation & escviption fa object vecognition pet Tmage acquisition a ienage Processigg. pan “genes There ave two categories of image iovalved 1° the Methods whose ‘IP's @ olp's ave images .metheds \ahose ofp's are attributes extvacted Pram hose image Mi * Teeage Aquisition STE is acquiting the age that is alveady in 7 digital for that is digital iage = AP it 15 not in digital Porm we need to convert the image into >. digital form by using imoge processing : “Image acquistion stage involves preprocessing suc cd b as sealing ete ¢ Fry 2 2. Image Enhancement: Tt is a Re ais process of adjosti “Pe qepwating athe digital image. so that results tog. oF neal Fly ‘splay ex: Brightness contrast & adjustment . yemoval of fala ply For in palute as Some may like high satovation Z aay Tt is subjective images & some may like natural colours. BcTmage Restoration; It is a process of ve been deqraded. Tt fs coenpletely objectively in nature 41 Colour ienage processing: at is a stage where processing of an colon 3 ferages ave done. Tt handles the inage processing of colooy images | indered ionages oF GY inrages 3 5. Wavelet & onultivesolotion processing: wavelets ave Small eo aves of lieeited duvation which ave used to calculate wavelet -transtom which provides time & Prequency Information of an image > Wavelets leads to multiresolution processing cahere images ave VEPIE- 2 -sented in varios deqrces of sesolution. 6. Trage compression: Tt is a technique used For reducing storage D space required to save an image. Tf the image size is lage it xequ- “VES MOE Bandusidth -to display an ‘Mage Hence image compiession car aM used 10 seduce gandwidth's, th extracting ab image covery and image “that has ~ OR dtdeeded “ Leben ee a er ee Be ef fer Bus % a. Morphological image processing: Tt deals wi & components that ave usefolin “yepresentation & description of Shapes. w Tt indodes basic morphological operations ke evosione dilation. iaq ® s.Tmage seqmentation Tt isa process tel of partition oF an image 5 ito cnultiple seqwents. TH ls used 4o locate dbiect & tooondaries 24? io an object. in object Ttisa process that assigns label toa te consisting of a cav this page should cea eakich identifies a Carls FAT? a-obiect Recogttin: “7 9 saa image. Ex, TP mage PA] D> process image assigning o label — 3 present in the image. vod ay 10. knowole dge toas¢ knowledge aboot Pt Cd ay piace ssing in the foun ob knowledge ba | 5 elated to iiemage processing 15 present oblem domain 3s coded in image se. 60 all the information in knowtstdqe base- Dxthe olp of the iraqe sens0¥S axe continuous in nature 2 a : 5 7 a digital Processing image must be in digital Porm. so we need < fo convert continous sit intd digital form. For this we are havin 4 = td main proceses :a } Sampling 5, quantization : 1. Sampling digi alizing “the spatial coordinate valoes is nothing bat e sampling / : ; 2.Qvantization: Digitizing the amplitude values of the image is known & as Quantization 1 =) 5 =p ee =" tet us take a line ‘thwugh a image in ordey to understand the concept of sampling & quantization. we move from point A toB. # sampling’ To understand the qraph pw pon and digitize the image, we will \ | divide emtive system into equal parts VU Dividing the image into equal parts or coordinate valves into equal parts {1111 t)i sy) is called sampling noaba ei eet eet tats ss # Relationship bl pitels. > cosidey an imaqe #(%4) consists of some pirel valves have shown. in below fq where P&G are particday pinels “2 coe ¥ A quoup oP pinels 1s called as subset itp is denoted by s- An image is represented as fellows wepts fo t ofyy : 7 4 af | wd weg [aang 2 Ll] wt |g ae evel | wha) pele il 4g * Relationship bla pixels ts classified into 9 types ait Neighbornhood peels Cor) 4 neighbours TP a pixel pOay) fo survoonded PY two vertical & 400 hovizonk | hood pitels DS-al labors than it is ate Netqnbout denoted by “the syrnbal Nergbboos 7 " eit is a me NYP) = tages (4) iy), dd) 3 why | ed | wig | uyl sospect © 2 - aM 7 2. Diagonal neighbours Nol P) * 4h a pirels From fosm A with the middle pirel {t (5 called as the diagonal Neighbour 2 — —— +h 7a Not?) = cary) Dd ED % 16 a combination of Neighbours and @ 3, s- Neighbour Nel°) + 7 @ diagonal neighbours yet ga) Cat g) Re rr et ft at at ll ve bv Vo dvoOP REC OHOES Ny (PD + HOO) a4 > et yy, | niglP) = 4. > ay feted oy a) path Pe Jett [en | Er > £ connectivity [Adjancencd are neiqnbours and hexe some gray level value ay w TP two pitels -that + badjancercy taken be connected > a* comectinity is the most important io the image process t0 connect the pirelS i. hey ave classified into 3 types en F 1,y- adjancercyGira v + Binavy Tmaqe oot . A jancenc oo 8 adi 4 # oo Wo 00 Not connee ted @) Not adjancey 5. M- Adjanencey Tt is also known as onited ° i) (—— “ 1o 2 adjai FTE is used 40 eliminate “thre ambtitoty Cai adjancency usage att says that we should give top bot not the diagonal connect x Distance MeasuTe > conside¥ ar a5 shown jn below go) ] @ Not connected ov NO image £041 y level Treaqe - £1203 100% 5H 10 _ 00,5 81 150 27 34 ol 200 3 4B zp ME 56 + adjancent ncency 7 € t PPicolty) that occUTs doving 8 © tT priosity to straight tine connect @ E 4) having thtee pitel values & < FEF, Piqié = particolar pirels po 12 5 . + 4 z "ztayb) ® properties: LocPq)eo 2. D( Pq) =0 4 P= ke DCE Pp) ¥ Avithemartic operations blu images bsQtg)= fu y)t qiy) 2.b (Uy) = FOG 4) = 404) 3. pOiry)= $004)* 404) 4. dug) =? Ong) A 9004) aa eI et Span ee eet oss Fee