Unit -1 Tntiodyction to MATLAB © Unit -0 s , Introduction to Drip System a. E et of I i9 to mage Processing Sustemm S pefiial Image: Repeetenintiinassics| Se i>. 33 oriti Data

[Digplay ' t 7 D/A ' Converter ‘ fig. Cleme : ots of { gital Im Pracessin System “ ti ¢j : ; A hree dimensional images -in MATLAR. With the help foreod () function, the User con read Sis tpes | ditrerent file formats as follaw ‘ Format Name Discription Extension Q TLEF Tagged Image tiff or tif File Formnat BrP | Windawws Bitmap file: |. bmp SPEsyntax : yornorne = imreod (‘filename - ext’); Ex 25> x= imreod (‘cameroman-tif’); On the imagers O eod, OFter proce ing it Should be saved in ent dictionar of MATLAB colled work directory» Tf the files ones om (0 directory 1a Necessity O ‘ting th Complete path ile. However, “if the images Ore ftom Outside the work director ompulsory to os e noth ok Fhe file qlong | with its extension - : A Displaying snages si Once the aa Gre read. in Order to a eee imoges on Screen MATLAR uses Q ion colied imshou nction: herefore eacbnuics is OG predefined Function D ed fo disploy oll tynes o e forcnar that | pe Can tead 1th the help df imread ochions Therefore , with the help of imshowlTo write the text MATLAB Uses fruriteC) le “The syntax of imurite C) Function is Os follows + Bl Syntax : >> imurite Cvornome . text’); Where, vatname is the im i sist The text should be enclosed in a pair of single quote » For Ex. >>dmurite (xs'Man’); The Ohove Statement coil! uorite the text Colled ‘MGM’ Saami adiite (1) 'MGr’, [ Po,100) ); erent will display the text ton Called called $MGM' on Specific Fr land column position Called 100 - Jn this woy pine can ulrite the ted | he image with the tle of imusrit a function - I | * Tmage Types : | ‘ as Qre different tune of imag ed r 5 : 7 ; S work: These ore oS below : @) “Manochrome imag @ Binary image @) RGR images Indexed images Monochrome Images: Monochtome images are the inte images in which Puvety pixel ig_fepresented by 0 brightn Q olled intensi alue or gro e - The monachrome images Ore o Colled a Intensi ightness images or gra e imag To store such of imaq Q comp Yee System requires 3 hits. erefore, Sometimes these mag Qre Olson called os 8b images - ahLLL << — +} ingry Im eet £ 7 images hove o very specific Beco on - ene fohieh ratte ae tio 4 represents cohite colar -To stare the binary 1 it ea a pixel depth of binary imoge i Mee bike 3) RGB Image. a Bssin J The RGB images ore the colar images: re the three dimensional images in =—_}“These images a ‘ eA piel (c_foomed ty te A Tt r G I i colled of 24 hit mages. y Indexed Amages | Oe of the drowbnck of RGB i t Lit tequires huge memary for ae To duercomne this drowhork , the eztended versionpune far eoch af the Component is reffered a g dentificotion Ntirmber Called Os ind nD Ekie 2 these index numb Qre maintat o foble colled Indes Toble « r Oo Im oges Into indexed images and indexed imo B image ATLAR provid 1o pre-defined funchions as: @RGB2IND : Converts RGR icnages ta indexed images @ IND2RGE: Conve I d imag O 8 imag —T Fig: Representation “Of ‘indexed images| Data Classes : the user is Storing in vorioble - Like i C, c++ Vorigbles , MATLAB 1ses the Concept of donee, | 1 for_performing — : : b | indexed image etc: MATLAB supports three a L ae classes as 1) Hamertec! doto closs 2) Charoctet data clas eee | umerical Dato Class: j A Numertcol dota class ist Sed For performing all eer A pumerical dota class i hel divided into six types Os @_UINT? : igned integer ? dato cla al which requires @ bits “or 1 h efor | orag oinge @) VINTIE : Tt requires 1é bits oF 2 bytes for its @_UINT32_: Ft requ! 6s 32 bits or4 bytes Fo Storage - @_INT 2 : It is 0 Signed integer & data ‘clo bi ires 2 bits or't byte For its Storage : |@ InTic: Ti tequires 4¢bits or 2bytes for its Storage —_ aT On @ INT32 + Tt requires 32 hits oF 4 hytes for its 2» Character Data Class +, [hea Papi e ty aietaidat Eoin ae TP paca ng Numerical éalculqtions an _ ima ay the data closses_ore sed e USer js ! Storing. * Introduction to M=Function Prog ming: nail eedeaeioe ee 6= cel | [he advantag a NQ =a ned haere aes d pction avoids the tepeqtal sea: ble of mfartah The MATLAB function programm ing - Fallows Following Leps fo corite “9 Function as ¢ 1) Function definition line I +) Hi Line 3) HelpTes t » The function body Oommen 4) Function Definition Lines Zt_is pees but hae ae Ned D In TIATLAR fe Q ored ina Q eSe a ord a neaiektings 4 Syotox > >> Function Cautput) = nor Cinput)z J Sth ¢function-name olonys beqinsiith a | i 1 Chorocter or alphabet - It r Sumbals_but t L ngth of function Name is of 163 choracters . : a @) Spaces ore Not allowed in function Dome. = Ltt LING s+ as] Tt is the first text Jine-This ig the'header | Tine which is generally catitten of th af the ragram Hi Line is the non- executable statement 3) Help Text : as ; : all Dan-executoble statements- L information about ¥ Lp| 0 oction Name Bus = >> help imread () 4) The Function Body : Tt consist of Q set af oll execytob Se Stn ec - Shi he important blo hich —____iconsist of the o Q pO e_lagic of 0 ) sanniio dilator a written ‘ip function hody « BPRCGININe Nese: on ommeNis ore Non =ex QD ments: The co, ill Mot pro of Non- executable Statement . Simpl. 22 | program «eo 0 0 na Q or the i Onvinience ore ‘ting th Comment jp 9_pri oe In. A the“Cammenh peg mb Ine f+ mbo is provide: Mitel ek pea oreen Calor Tao will be considered q eee ot line in aL f Tn this Way, the Licer defined funch 0 D ATLAR.| three types of dota classes as 4) “Numeric. data class. ay Chifiricter dotarfiag’ 3) Jiagical data class Gill Poohintbreina ‘ithogesi ting yao vdetoult dato closs os WNT? «In MATLAB: Occording to the User's requirement, we con. convert from _one data cl ass ta Onottier data cioss- Conversion af one dota class to __ nother is called Converting between one dato closs fo another - To convert one data class to anather, the — ——| following s7niox isu : Syntax: >> vor Name =newdataciassname image iro S Ex. X = imreod C cameraman -tif')§ oh ar ing data class os [INT ® . niep_thi Q D closs fo ONoiney Goro ; eguitement - ¥ Digital Image Representation : oy: i is defined as a tuo-dien = QO ere p A fou) positar piel icici ate value colled Giay [evel or nten b pisel- Ip order to icnplement image pro ing 55 = | into digital form ‘Therefo on ion of anala: image to digital fore alled image digitization or a_sampling process: Digita) imo ar p ated ip @ Computer Ferm using tino Methods Os : ga aa Convention System Y Matrix Represeniotion 1) Co-ordinate Convension System‘! Hiso simple co-cidinnte syste po_whiel is ne to lege O pixel posit! 0 througho co-ordit Qs %* These Co-cudi notes ore. " : -Bitseine. Bt { (050), (01), (0,2), (0,3), (0.4)... (0,M-1) oni the seiait ich such 1as C130)> Ghat). CAi2) 5 tan | (a M1) ‘Likewise. the Samples af et Ore located on the Co-ordinate (9-510). (m-1,4), ($4-1,2) ian “The-obove complete Sampling process con he represen : ee ; cols > ; 1 nn 2.433 oon uNal - + tt ok one 0! L + oi ! fi } _|| Rows E — ne : ce Nt)2) Matrix Representation = | the matrix representation - Ahenever the imag lore stored in 0 variable, such images ore Stored | in ee wi sais en ae computer. matrix lepteceniahion of di ital i images eat e tepresenfotion of a pV et US Consider an image ho ‘090 ize This image is teotesented in tHe form of matrix inside the memory 6, Computer mM ample is f (0,0)~and the lost gomno e is 3 M-4)N-3) : Sometimes. amp Q Q im en} picture eleme nixe cae ik of meniio xel 6 The reoresentation of dinita images ip 1d, Mmotrix is of follows : x4) =[$ (0,0), f(0,1),.... F(ona 1,0), f(t thea f.1), ~f (2,)N-4)) J oom (hs).f faa aie eaUNIT -JIr Thtensity Transformation £ Spatiq) _ (rer | | ifterin | ae Background: : em iter no id action h_0 O DFT (Discret fer Transform T(Weblet Tro 1m Distt (Discrete Weblet Tien form) etc: then that type o Q Pd q cy damain on he ep ented D equation as: sy ET Cf Goy)) ae 2 (x, epresen he inp tL image h ansfarmathan Nction a equen gamoain process1n9 e i af image processing om Da ‘complex to Understand alsa difril A o Lind and _olsa d oO implenih gray levels fram Ota 25° Can be shaun H (rk) = 1k a “Dk! Feptes ents the iad bres tk gray level = a throughout the image represen ts_histogram o} the image. 7 For Ex : H(10) =100Som#times the images ore moderate i Fn dorker i Q-In brig tet: images the gra uc ee 255 and i moderate | fr y levelg Qre _ i i ram MAT. i imbist C) function. Syntax : >, imbist (var name) j Where,—<—=— ated L de >> clear al) i >> close all ae ( anan-tif'): >> j= imread:('Comer: [55 Joure ; Syl ow (i) 5 = [> figure icohist Gi : OT Piaiinistooray which shows the fepeatation of Sp cifi > Het inn | / ising bi fogram of the impoge: Ry 2 bieaeace can be-disoloyed ee ’ tthe histograrn can be displayed : play Fepresented hy Using 8 tupes Qs J Bor histag ram 2 is r S isto:[eee Te | tal ford oO plo q D Q 0 mandato ID are Usters a Th q Nothing b the group [Buds should be Cunfed thin 8 ange c} 0-36 Ihe minimum gray le of _o pixel is 0 and H maximum groy level oF the pia js asc Within his range, o us Q Q No-o £ geth Depending on the’ 's tequiremmen Need th 0-0 bb Crom tet Aired Bei pees i In det tn plot different types of histogram as bor histograms Sten histogram ond plat histoqra 3 _pre- j i as? hor ©) %:8tem ¢> and plot.o) >>/- Program tn plat peso histogrann. > ; Cleat oll i >> Clase oll; >> f= imread ( heed inci tip’); L » fi ure} dy how Ci): hte >» igure ; > Imhist Ci); >> jsi(1:ioirse); >> Figure ; >» bor >»> Figure ; ae 2 Uses i rapheQ —_—_ eontrast images* In tive inbage thvrdliale Lt lot aa or ond the User Connat ident (Therefore in_oider to observe the details oh it js Necessary ty remove corknes or to. increase the bri if the pial Thi tosk @ =Gtion functions - Once the inten iy Fronsformatan Pay gia Ore Opplied on the images, the intensitgs pixel gets acliusied Qutomatically and the ines Vevel af the pixel is increased: The s became more clearer ty understand and t oe con fo 7 ype aod exirodt a nglow in’ hi h.in) represent i t param Clow. out high-outy! tet porate! He in } the ret 0 2 Sr be oauridd ty EeSiang 6} 90-400. [ Sorneticnes the intensity tronsforrmation nctions Ore olso Used tn abla | ‘bicin the complement Ll. mnonachtom + image. Todo thi: simodjustC) i pee be used of eld = = fonadiu t (i, Co 1) 04.03); “The Syntax of imcormpleMmentC) i Gs follows > >> var name= imcomplernent (image -artay i2 >>> clear alls >> close gits 7 >> Tsimread (cameraman - tif’) 5 > figure § >> imshow () ; é » x= imadjust (fT, C01), Cr0J); 7? fi ure >> _imshow 0); image Using imcamplement ©) function 4 2? Clear aii 5 | >> close ali} : >) f= imread (cameraman tif) ; >> figure 5 >> iMshow (i); : de fe imcarnplement (i); >> figure : >> IM show fi)al = J ti teri nea : : , ee mage piocessing work such os, image i j ition ete. 7 ‘The i ros fe naps ad Makes the eth More at fileding pice at he easily identified Pxtract from he image : “The Spatial filkexing techni ne es ate brood 1 ed. into two fi pts ae: A 4 Ot Spatial Filtey 2» Non- Jineor spatial Filter 1) Linear Spatig! Filter of ft the time of applying linear Spatial | iter, Firstly the Ltighhourhood oF d pire) is defined. A Neighbourhood Is Mathing but the collec r group Ixels ised for ptacessing the Images cDifining 2 atighhauhaed ts hat Faliduaiing tages as”: Se. f Ose. ee.een ey 4 (oe! aa) * inn ges Color images ete: The neighbourhonel size i: Ses EXS | F.xF 7 S53, neighbour ood wi 5x5 Neighbouthood _ eens generally 9 process 9 pixels ato time { She example cf $3 Neighbourhood is g Follows Oct y-3) PCy) fF (x-4, 741) Fi Me pa yt) f(t 4 1 (fH) ytd) ay The examples ay linear spotial filters Gre a4 follows : 1) q) erage 2) disk 3s Gaussian 4 cian ) fol hI 9) L0G (Laplaci i T cael plocion. oF gaussian) | _7) _prewitt iL 9) Mation 3) £The linear snatio! Filte Ore crea ing Q pre-defined Oction colled pecia funchen Cnc the fi Ite are CGied :T> Opp | MATLAB N por Q pted ined Q Hon | Colled jmfilter() Function. ,follawing is the syntns of imfilter ©) Function ! Syntax: >> varname = imfi Iter (image =0noy Filter.a1rop, ear image. =Qttay tepresents the voriihialedhieh | f-the input ‘image. filter? Qitoy tepresenis the vorioble which stores | the filter. cf lineot Spotia| filters Oreos Fallows : 7 a 1 1) Average : —_}Syntax : >> Varnome= Fspecial Coverage’ Cee filter af mh ail e e Fulbet js_ 3x3. DX 5, AK7 9x9, stalaa ete.tObove syniox will define o gaussian Size rx § the1 . poo Laplacian of Gqussign —_ SuptaI>> Vorname = pecial (‘Loq', [+c], sigma, alpha): -_ | [he Qbove Suntox will d ine the Ioplacion je ia pe lone titers te papi the param bh a | Q Neighbourhood size andatd ‘deviahon called lama. “nnd alpha. B d ! ATLAB will Qssian 33 -Deighhourhood: ‘ond 25 Sigtha and alpha c Value for Log friter. 6) TMotion : Otay | >> varname = fspecia! C'motion', theta): #5 Plienhoie Yntonx will define the Filter Called Motion filter: aU sccanrne I as—thela. This filter will rotate the image in a {clockwise direction or in opti -clockwice ditechon: lc, positive theto yoiue wifi Fotote the j in__clock wise direction and the pegative theta Volue Ul _totote the imoge in anti-clockwise ditecHan: ppt hue synins will create a filter prewitt This filler is the example of detection Operatoy which is used to| [recone (oot | Be : a 4 | Premith Onerotor Js d_to determi Vertical edges af the imoge-It does pal equire Qny Oroumeni CF paramerer~ 8) Sobel * Syntax: >> Var name = fspeciol ('schel’) > The pove Syntax i reate a filter it js qlso one af the exomoll Called Sobel filter. Sh) _berix edge detection operator which is used _determine hari onto] edge: of the image. It does nat Fequire Ghy porameter at argument. gy Unsharp petyotaNs >> vornome = f: pecial (3 unshorp') 5 “The obove suntor will create o filter Called lunshorp Fil Itev. This filtey js USed-toy make | nsharp: It also: does equire an argumést or_para meter:Page No, ‘Onto Non-linear Spatial Filter : A Comman d-foo) for removing n-lineay patig t Sometimes, Non-linear snatig e Q Qo Colled a anking filter. Sip Q pon-linea anal filter ess Ca, tation time and ince Such filters Qre easy ta implement, the nov - a | = | Laci € ond m the images is then [ae | = | ra inea patio! filters ar to linear spatial filters . ‘There ate two { pes_of nan-lineat Spatial) filters as : 1) Oder Statistics os *) Medion filter a 4) Order Statistics Filter : : The_order Statistics Filtey is the non-linear | oatia! filter whicl j j depending on the arder volue- Since’, it implement | depending on the atdet value, hence ihe dine larder Stotistice filter. To ienplenn ent this’ Filter [ marine ~ pporls o pre-defined Function called — ‘ord Filt2" 5 Syntax : >> varname =ordfilt2 (image onay, order be)), | [ | image o epres en rhe NOMAD hich | | On ncicd } Ord Sear Hoa eae fies eIn_the tang of 4 te ae minis ondeGin fafisties filt hos Th. i ood Size. ; 7 j 3. default ighbo rhood Size is 3x at lever, we Can “change the neighbourhood 5X5 FRI, 9x9, 11x14 etc. 2 edian Filter : “The median filter fs first dit French mothematicion called Medign.Henee Dome is Medion filter: This filter is g; ia the Jel OS ‘Color image- nelly. there ate four ypes of Noite asi D poise 3 ise EDS Speck ¢ Daise 4) Poisson noise oe medion Filter iD Prat, led =medfilt2 Cimage arra 2£ # Write short pote on imadjust C) - | _imadjust C)_: {inadjust ¢ z ay fab func MATLAB oan o_pre-defined function called as The syntax far pede function is as Fallows ¢ Syntax: >> Vvarname = i ust imag e-anray, Llow.in eee _out high out) ba a variable which Where. Image airay leptesents Consist of the input image low in hi hain’ bs iSpeciglC) =: ‘The linear Spatial filters are created by iP & i i | ed. ig C Once the filters ore created, ta Qe the ters, rts _q_ pre-defined function Called lengfister C) function - Ha i if x4 ial ©) Tpernng isthe syntar far Fspecial ©.. ansist af : 4 ilter_ Qrr ts th | ic E maces ieee son ‘alith the help of imfilter© function in _ MATLAB. different Lypes of fillers Can be dpplied— ‘the image for it pracess Dg is neighbourhood ? Explain u ith Fxamale xel_is defined - Pixels used fay plocessing the image. * Advantes eo aks== —=—<——== = = a — gh) bo NGOd per ~ ee, Us ce Or |. The Neighbo Oad Size sf fy, ete. 3x3 eiahhovithoad t Process Sipinels at atime: : a 5*5_ neighbourhood ill ixels at o time: generally of _ 3x3, 5x s process 25 No-d}, On Ex «: fOr1y-s) f(xi1,4) £G-4, yt) (ar421) =f Cay) 6b, gH) Cut 14-1) "£(a44y) £ (+4, yt4) ins neigh houthaadfe Unit -I Introdyction To MATLAB x Advantages of MATLAB; ATLA = a‘ Other seantieat langunges Si eh AeA. eee Ore the importnot odva using MATLAR a 1) Eose of use 2. Platform independent _ t 3) Pre-defined functians Groohical Uiser interface MATLAB compiler 1) Ease Of use:So Qre. MATLAB is Supported erating Hers Such as lAlINDOWS . UNIX, LINUx , ete. [he proe ram 2M ON ane operatin em an be eq ansfe ea 0 ONoMer Operating a : q Systecn becciuse itis ¢ platform independen Softwore. 3)_Pre-defined Function: The main advantage Of using MATLAR i the use of hy pe [tbtary of pre-defined functions. f PIATLAR Uses total &e toolboxes. Same of the lbnxes Ore ZPT (Image Proce ing Toolbox) , Neural network foolhox fuzzy Systems ete, Loch of the toolbox has’ minipauen 1000 ore- deFing fuoctions Therefore, MATLAR ho minimpucn 32000 pre-defined Functions. Due to Support of bupe ple-defined functions writin Progracn is “ety Simple as Compared to fest f ptng-racnmning languages. Device Independent Pio ing: Unlike most of th ther o ogramming lon ig ta display tmaae guages MIATLAB O Separate indows call e“windo idle MoOtonN €Due to Support Ok GUT ony inespetienced Uber can easily work IN MATLAR- 2 6) MATLAB Compiler : sal J MATLAB ’s 1 exibility is to Work with PIATLAB Compiler. MATLAB Uses compiler . translator - a ‘ompiler will read Seeramplte i. Completely at a time and canvert it _inta machine ; lewel | Therefore, _the conn i ig very Foster in MATLAB oO. compared erefore it is O very expensive Software ond — erally used in R and D Section Gh, Companies .2) Comemnand Window i 3) fipiemons History Window 4) Editor sy ost ‘Window Pe ce Browser ?) elp Browser *) Path Browser 1) Start Butto As Soon os the MATLAB So ove i Q MATLAB desktop Shows the opeo apenin button of MATLAB is called Pe i aa All Major tool ond windows ale Can easily Acer from sfort button: The windows such os Compan window, command history window, fig ipdou olay editor etc can he “easily acce 1. a the Stort button. This butlon ic similar fo #1 button of windows OS: 2) Command indo tight hand Side of default MATLAB De Command uw -*Thnd the time on which the commands aue _dauible-click an that comenned, the commands will he 1 €xecuted- I ‘yy Brgy Editor : F J to_wurite. 0 program, MATLAB ___ tor colled Quay editor. The tomatically Os soon 05 Q Use Otay editor is Goen a a. Ser I creates a new [MATLAB file» Once the i - pest for each of the line oe MATLAB Compiler O55 19 Nig i Fic Oumber Called a Oumber. This line Nuoiber Con be rererred in Qnother pease e Orroy editor is the pragromenin editoy with MATLAB anguage features fc different colors: The et edijor always a fa eD Ft displa eae BS fe nocdin olor. Ta omplete character data . | 0 onthe program il display jn re : | Built dn Ord nO e olor. A Por inee a: | us DIN Q ptoaram | SE igure Window: Bests. ug A Figure windows is - eu to display MATLAB arophics. The ep dh pndow Q display 2D ¢ 3D jmooes. We' on also displa : : ctr. 0 fipderl Qeo Onimatioa Dra ph, plo chien Gre tuiddow- As Soon Os 4 fioure command i$ given OQ MATLAR compiler displays the Separate uNnda allied figure window - is_ window displays a enu rot the to undow- Tt consi Various menus Sich Qs File, edit, view, insert tools , window help ete. each of the menu ha Various menu aptions A_user can also mh tos, X-oris lobe! Y-axis lobel, heod footer information in different colors Using Pye Wwindous, € lor a Browser: — LDP OMQbles reaps “oh LlorkSpace browser tmindou is inf roam + Q3 Soon as th qriobles are crear af the oennary esetved for the d of VOriahle is to Store a data: i Size or th bein_ Ar with ol! ohove information, the Workspace browser also displays the Spatial co-ardin i each of the pixel ao it’s inbensity value. All Such deta We Can easily access Fromm the sinc colled Wor kspace browser z 7) Help Browser: The help browser Window is used 40 Obinip the hejp any pres defined function Com enands ete. There Ore three ways htain the help in MATLAB os : 1} Using help browser ii) by tising helo command Ee iii) by using lookfar command miilsi S s ATLA i brouser (?) is displayed. at the end of IComenu bor: By clickin browser jcon, the list of oll pre-defined” Functions will be displayed. By clicking ana Specific pclae the help —| pte-defindy function js displayed in commond Window: The Second way to obtain he help is by Using — help command. aBrowser : path browser windou directory or fo MATLAB wi e Ql) programming Current delay eatied Wark directo eaeual t the oo of users directory b ing PO 0 CTh h shrews ef Window can he apene Using Set path option | ftorp file Menu. Once g Jo ste th th ing Save th browser. Nox Ore Fe these pla th Set aaa path browser all _obove en ATLAR nN Environment. =MATLAB Supports tw s Pa 4) Arca ration 2) Matrix operation iN ) Array Operatign : Zs i ich is tao s ent hy elemme i y eration 13 called Ouray operofion - This | ig performed with recpective elemen between fwo Ouoys. There Ore Four types of Fundamental quay Operations Qs : - i) Ptroy Additian ii) roy sublrocton m4 Aras ulti plication a iv) aut DivisionArra Muyltiplicgtion d Auroy coultin qtion is very different 0 Compared ae penltp ation [o_perforn Special symbal called + (dot) symbal- ote On=. CO . eb iv) Array Division : Array division is Qlso very different from matrix division. a6 perform ar Loy division, a MATLAB: Compiler lises 0 Special Symbo allied + (dot cob [Mmexeec= y./ 5 2> Matrix Operation : operatians Os !i) Matrix Addition ii) Motrix Subtraction iii 1K Multiplication iv) Matrix Division Some of the common an Ore Qs faliows : en Operation MATLAB Form. Cornments Aroy Addition ih eke i cnatox addition ate identical. _Attay Subtraction deb 2 Alay ubt od L trix Subtrachon identical. fray Multiplication Q-*b Aiay couttialtcati will be perfarmed by —Lelennen t by eleme: | basis het” two O'rays calied Q and b- Matrix Multiplication Q#b {I - ji leshasis - Thal © [Matis Division | a/b Maitix division ‘be rees 1S 4 fen 1 : determined b QD the following formilq Gist ils g_* inv Cb) Tn this way the scola Qnd otra i fe petformed in MATLAB - Since in MATLAB the imaged Ore _tead ip Matrix therefore Offix Qperorians are fh gu tly used_in innoge ee (wor. trays: d The fundomental unit of data in ony MATAR program is the atray. An any j the Collection o ae data values having’similar data type. Tn Amy. he er COM only Store the data wales “of Similay dototupe- The Dis of Oalied Sents maximum how many no 4, on Stare in the Orroy» ‘The Orra’ Q script: AO Oray makes the xible and ea d_und and - Qus Ore wide ed in image pro in Attays are cnainly Classitied into three tupes o 4) One dimensional arto. 2 Tivo dimensional Orroy 3) Tluiti dimensional onioy 4) One Dimensional Array “The airny Which consist of Only one dimensian | Upe of Orroy is call ee iaoo! eat. ie dim ens'onad ones 2) Two Dimensiong! Atray : The Qtiny which’ is defined ‘by using two imensions os tow and columns such type of Sun is called Os two dimensional a ar do hieetnss dimensigna! 2104 - Matrix_is the phost comman i is _qls | ~di i ide i 0 i Ww j i nd column « Ex: Of2)}[2) = $1,2,3,4 Yowie lelement ore Stared [n column ‘wire i-e- First c firstly, Second column Second ji i is the | f Vey nsist of} three dime Gireen_and Blue- “The pixels Of Color i. : Eninimum three voliies which bepresents the Joteositie | three dimrnsions as Reds Green and Glue. | Ex: q [123 456 389 101112] The Size of hove matrix is x3 which consis} dh Mi tows 4 3 colummps Therefore, total no elements #43 41 94.1) 4 q (21) Mins (.3)1,) Jog (4,1) 429 (1,2) 50 (2:2) 8a (3,2) 44a (4,2) 3q (1,3)ig Memo epresentalian of m -dimeNsiona atra ing Column cnajor MATLAB as a Scratchpad _: MATLAB is mottix Jghorty Software hich d_for perfo oll aleuJadi M. is _Us Qs digi i e processing, dig) igna! Meurol fiers 2 fuzzy nce: ete Tp it’s simplest form, MATLAR Sometimes Con he Used os @ Scrath pad to perform Mathematical calculations; A Scratchpad — 18 Simply o calculator which is used to perform “the f mple Operation b ti Ss Ho. Multiplication division ,ete Soon OS MATLAB Sot oft wore is Opened O MATLAB Screen -is-divided into thter windows as follows: 1) Command Window _ 2> Command hittory windous j 3) Current directory window Among these three windoms, a command Window will oct ay Scratchpad. in MATLAR. Tp comman Window, o User can clinectly execute ali pre-defined octions and commands : To perform simple calcula =lons like addition, Subtraction ete no necessity oh iti program - Lng th als Such od y=) # fete a User J - Con_perform Simple aperatons os addition, btraction etc.d co Qted and displayed to the user: B Dg. f ——_—1 Same Standard formula a andi, de oon ———Orea dh circle, Oreq dh square, volume xf. ind etc No Necessity a Hing Q program fo hb type af Aperoticns « L : Tn this vw) ing direct formulae | a urer can determine “the resuilk: Therefate s ——_|Sometimer MATLAB o as _@ Scratchpad or Simp 4 Calculotor- * MATLAB pe ——_—______ Operator Save th sign axsy mbol Used to perform Operations. Oper nds a arigbles an 3 which the operation takes pla p___MATLAR operotore orenin ified into thr |_types ase: ——} tithesatea! Operato rotors oie ‘Logical Operators as: 4) Authmatica’ weeratu G earithmatical opera S$ gre th Pilabhenteain| or fundamental operatars whi are Ufed to perform ari addition, SUbtrOction » mM exponentatio alculaHo Q Iplicabion , division Follawing Ore the €xOmoles of Qrithmaht 5—— aoa = = canal Operator: ‘Operation d rb Addition e = Subtraction : * Multiplication g A: Division S a Ex ponentiation 2> Relatigng! Operators : ! ‘The op i s relational fxptessian or canditianad exptession, Stich u Ore Cc ed re tonal Operators - These Operators are generally Used jp ¢ if -Plse een to_evaluote the conditiane. following Qre the £xomples dh relational aperators. Operator Operation iz Less than > Greater than less than or equal to Greater than-or equal to is equal to Nes mole Nob equal to 3} Logical o perofors : | The operators which ote used 40 combine eet ee pene i cnerettical—Logical AND Logical oR Logical NoTUnit -VL ye __Jmage _| Proce ssing g b Color is:very helo in obj dentification and €rtroaction from?the imag Ti ier i echnig Ole implemented on the again unayémonly a With the gdvanced. te aq is Now pa Work on claw images alsa he Number of bi Store the images ore called oe e_pinel depth sis 8 bit whereas the pixel depth Of calor imagee js 24 hit | ctl al The Calor image processing lis divided inte two major areas os 1) Full Color image ptocessing « = 2) Pseudo calor image processing . The processing af full calor imag e Color images 12 ite inages ct ald UU Calo eetIn 38 1666, Issacc Newton first disco tho hen Q beam a4, alight pa Lnrough Q prism the emerging bh of light is not q white pemclicht bit; ansist of Sven different colors of varvfing intensities. AU these colors forms a 1 un: ‘The color spectrum) MCOnsict at Seven colors Os Wialet (v) Indigo (2), Blue(aY: Green (6); Yo liow (1): Orang eo), Red (R Jn short pitis termed as ViBGyoR.. This color Spectium is-Used for Calor ima, processing. > There-are different onethads ore techniduer 1_feptesentation af colar Spectrum . Thes e technig Pore called color models ar co paces. The representation o' t f color Spectrum Using VIBHYOR IS _Q¢ follows :‘There Ore three colar models used for color inage pigcessing Os 2: <5 ¥ ; PMA RGR ccolormodel- dod oR 2) HSV\ colarimodel 3) Cmy' color model 4) RGB Color Nodel_: __Tt is the fundamental > mshaw Ci); Be ued adiGs, sits >>» figure pao >>. inishow (i); > Vidneri: Gi 242), | + >» figure, as >» imshow (it); -_ | SSSR » afi ure, ; >» _imshow CJ2); #__»> + Prooram >> clegr git; >> Close Qu: >> is imread (‘cameraman . tip’) >» figures » show (i) ; » Js imadjust (i, fo iugshthol) 3 >> fi ure, a» IM show (i);is imread (‘cameroman:tif’); igure, ‘ fo Gi) « J= imcomplement (1); igUres inpshow W@W)| Unit ST Image Restoration Tozoic i 5 i hnigues gre gic images etc. Image restaratan tec Used jo remove the Noise from texiure images, mozai s. Notural Scene images etc: These techniques are BU sce fs monochrame imnages Qs well as color in ages # A Model of image degradation /restoration process Genera y the restoration plocess cansist of = Image degradatian is O process at addition o | Doise on the image to form Noisy images. Eom Restoratian is 4 plocecs af removal of Noise ftom the i es-CTherefore, to quark with the } restoration techniques, these {wo procester Ore ~Alecessary .ing -allabove de grodatish Junclians-0-uier— Ve cea different + ee of a tole gn Dce G noise i 1p various cca tie called Restate, aa Restototion tenes There Ore two t ypes Of testorntion Filters as : i asrear Restoration filter i “iter imoge_is_called Restoration Bho “tas hae —_ degradation and image testorotio are ——| to obtain better quali Output. J —— a7Os Haeeo it Guy) Degradation Res {aration} Gg Function (H iter ' vir : i p(ey) | Noise eqradation i Restoration 7 He: A_madelof image degradation /testaration I plocess. ee lg Noise Models: f iD > =i ise (imagegrray, type? arguments); bihete. imagearray tépt i Shi ist the Input imag e- pe-> Tt teoresen Q typ fof Ddisiele ry al Vite yatguments > Tt indicates the Necessary parameters used ta Add a noise on the image + CThe aise madels ore categorized into four types as : i} Gaussian noise model ii) Salt & Peppet noise made! iti) ee onde! sities ae iv) Poissam onise modé) i) Gaussian Noise Model: This naise model is first discovered by o ici issian. Hence the nome Gioussian noise model: This naise madel will add_o MATLAB assign 0 mean and 0-1 varian noussian Pontes he Gausion nase the dense Ndisefoileulehing designed — called Tedian Filter. To. implement onedian filter o ee. HAMAR Supports a pte-defined Functi on_co eae j edfilt2’ By: >> x = imnoise (1, ‘salt S pepper’, 0-02); Nanochrame iman requires an in MATLAB g QA color ipo One paramete ign 0-04 yicigl Ee Oh he ane oath images ond the ¢ Dois generates 0 random naise caer mes it is alsa co i Qs artificial random poise code! : | For Ex: x= impose (z, “poisson') 5 # W-A-P- in MATiaR to demonstrate _naise models - >> I- Program to demonstrate poise models - Pare aClegrgli: >» close oll; Z > i= imread (comeroman:ti¢’)s > Figure. » ffm how (i) 5 >» f= Imnaise (i. gaussign’,d-1,0. 02); » Figure, > imshow (J); Bers [l= Benet. Salt 4 pepper’ 0: ops >» figure, =ee imshow (ut) : >> ds Beeline ik Speckle’, 0.09) a » ute,Opp Oo On ne mag D eH nosition as § iy). T) mation” functions can be tepresented by BFallawting mathematical equ ation as: ep INp mage A ng Othemotically ‘the geometri 9 C4) =T (¢Gby)) ey) Cres U ; | FCoy) represents the inpyt image- fepresents the fronsfor mation function gle p) represent the outoud innage.- # - é Ore thre nes a aNsformabhon | functions Qs :| gra P of pix Om one co-ardina a Anath O-ordinote g hole om ane position ta another = position an ion is Gpplicable for different nes Oo ienag Oriol irpag = Natural scené imag : An order to divide’ the image into different ectians, MATLAR plovides a pre-defined Function called _ imctop © Function . B sing this function O User can Select 0 specific section and can work sper the beqe ferment . 2) Rotgtion : Rokn ion_is alsa one of the geametnical tronsformation function which is Used to rotate the Tictin . i e through Specific Angle. To tatote the i rovides O ple-defin ed Uoction Called imratatec Syntax: he t img array , Method’ Ogle);—ty default. MaTuAg will Consider Nearest metho 3) Scalin, Is _tronsfarmation et on which is used ta enlarge the ia QUDO nothing b n INQ Or a OF Manachrome and colar image 0 Hangers pmb of epics 1 beg On ave Gectease the Size g innoges uitimoately increases “the Size dy the’ imoge+ Simila ea ixels Of image So that the i image Consict af less Do-® Thi I pizele which wail) decrease the sine Hin . [eanctinn oy ier ing _or decreasing an _be donth two i A$ Of Ng of alg et This bs i con b Or ti teqistiation method -_ thicmetiod ie aplbabtd for anc Jmoges = indout the Statistical difference between tue BA, istogram Equalization : 4 The histogiam i i a.-gtaphi col_tepresentation times g Specific roy ipuel tepeats Ge istagram of the image con Mathematical equation as: z Mis | ic_gray levels ‘Dk represents the no-of times rk gioy de fits no} the IMQges histogram is is_processed. Q_luser istingut Separate. OE backned 7 ara els arto make o histagtam mate distinguish ___ equalization is d. “I istogram qualizatio Qenero Applicable for hig ptrast 1m) ogram gualizotion 1s used To Increase The Conm a g Orde o red The GQtkness GF mag 0 Mal £ i lizak th ay levels more clear, hi togram equa akon iS lysed . jenerally, histoatom.e, or. - nbanceme ark: In Q | Step ich is d To enhance th 2 a Quality output Js abtained- Histoar alization i Alsa-itnprover the quality the f ge 0 ay TT t, Mm Li i be more cleared ond the User can easily Eideoties ond tecagnise the abject fron the imager | Therefore histogrann £9 alization is use a LEE Ey higher 4 a Somer sptocei sie ware Such as compression, Seacnenta i fecogoition pice Jan CQ as Tp_order to perfor hist equalizaHi MATLAB ono Q pre-defined D j _Histeg©) otic Syntax : » Vorname = histeg image QrraClear gli; Close all; [= inread (‘comeraman-tif’); Q pi. [show (1); gute, imhist (i); ge histeg(i); igure, ipphist (J);ener on be_pro 4 2) equen domain Auhenever the images are processed directl [using Spatial co-ordinate of a pix hen that type ete imoges is called Spatig p § } EE The eto" i no 0 pes of damains Os : 7 Meoioldemin “sto so. a , +— | Lb ing mathematica) Hansformation functian h. OS FT (Fourity Transform), DET (Disctete Fourier Tronsfon DeT (Discrete Casine Transform), WT (Weblet Transforre) , T (Discrete [web et Transform) etc then that type Gh Processing of the i Imogen is. called frequency domain. pocessing: CThe difference between patio! domain with Frequency domain is in terms af quality. Also the frequ ep sing j ot 6) Home ta obtain th Yesults- Therefore frequency domain praces ing is rately uted for imag D cessing Work. * Filtering in frequency Domain < Filtering is Q. oe Lited For temovint Noise and b in ftom th 7Opplying filtert _| techniques the if IMoges will be Mare clear afd a ing 0 a Q Hering nig na COM easily perform bigher jn ge Oroce oatiq! damain thY gee 2! r } 1) Ideal Filter : ; i r Ore determine ine LF carries the details of information. or denth of information. The onie'patt of the image can b heel LEc Components of the The high frequency oc pixels uthich alusay fies on edges or boundaries of te image - Since HEC component “are on the edges such if ‘pixel conies very TS infarmoticn - Depend ng fan Hi a oup of . ixels the idea filter is oppplied thie fil Fill Further classified into ‘two types as i‘ ip Low pass filtey i) High pass filter Low Pass Filter : Whenever_0 low pase filter is Sapliaaee the ‘ih it wi it is e - Thereforarth filter j is Ricaib ered. Sboth iter The purpose of Filter is to remove noise and feet a : , ul ng i 2 This filter ir dicated by leas dB 1 ngmemati an_coassign 1 order value fa [he imag ansi LEC Clow Frequen Ompanent) + cig fs butferwarth fill 0. 40 GYOUps OF O a3 ond Wec(High Frequency | peed this filter is Further cinssifi d s iY Low) pass Butterimarth Filter iN High pas Butterworth filter C anents, then th aorth filter Once 4h - Whenever this £1 y 4s td lowpass | is. pn will ‘ee p 6. a eep 0 low the itu of aaa Frequency compan el tt. Sand: Mulia vdith Filter cen report the image ar core 4 the image - eafheReer _this filter is Oppli utfewaarth fite-On00 rth Filter Once itis £ filter is called ae s Opplied if ilen, Eg.-High p h th se rpeittenath Ei iter g filter - }_2D-DFT L2D-Discrete- Foyrier Transform] Acneng-all_ mathematical tran formation é I itr Transforo) is_g basic transfarmatia Lion: his transform is first discovered by gq Freo I . Joseph Fa -Hene the nome : CThe basi principle of DFT j hat q D hich periadica eneaty it is-q ombination oh, a_p@rts Os!ert) D also in ambination o and_negati image is te-con dat obtained domain ta spa al domain ith = segere Hooss-of INfarmation » Due do this imoartant choro 3 a d to image pro ing work en Ma eae 2D-DET in MATAR : onsider the imag 24) having a ek ize [MNJ wohere " E097) 2559), Ro MA8 ea 031,25 a ren, N=d Rae fe denatvs, le eee tow) thich i abinined by using abt ppluv)= 205 25 a a a eee (frequency. t or positive cycle:bnique bos enhancemen pre n -Seqenentotion etc. Ta ‘obtain this firstly the Zen lpcints af o Signles are dete ~The Ze10_cross1 pci in deal alsa neg ponis Ore a OID Ticraane n op ortant charocteri DET is that whenever the image is reconstructed back fiom frequency domain tospqtial domain, we iL obtad oe =z tve the image os its with or loss of nformation. pls peity DET is-widely-used for develaning i prac tscing Applications. ie The representation of Zero crossing paints is tve Se arNS zero eds Sa Eig. . Representation of zero crossing points= Lh EP Apply FETshigt 5) Take Log ©} Apply PETShift - 2) Apply IFFT2 Stepi : In order to determine DFT, it ig m3; Deces: we [Step 2: i is i he image - “Jo do this MATLAB tises Q atead q 2D manaochtome image Fir Syntax + > vornome = imread (‘ filename. ext’); Ex: > i= imread (cameraman-tig’): i fonsForm 0 pre-defined fon Coll ie Syntax : >» vornorme + PPT2 Tinogeareal ! ee ne arnoen zs>> 1 =Qbs (x); Step 4: Once the absolute is taken, the hnanstourn i hi: do Toward pre-o Nea D Ie a Centre partc} the Image To-da this MATLAB tues tre. By shifting the transform Is comnpulsory to shift the transform lon called FT Shi nection. Syntax * >> varname= fftshift (image aay); GRD >y Yost prtshipt (x4) ; tep 5 : Once 0 low Fie ion is ——TIATLAB_Uses_log function —t. quency Components Qre determin I I starm.-CTo do thik Syntax: >> ee = log Cimage array); BBxaniin>> = log (44 abs (4 3iS (ftt2.0) function. | ipetshin © Function is the inverse of $Ftshift C) | function and ifft20 function is the inuerse of FFt2O oction - Syntax >>vorname = ifftshift (image array) ; pees Be ipeishipt Cy) Step #: Another inverse function fo transfer the image back from 9 cy domnin 10 Spatiol iol domain is LIfFt2C) function - : potas : >> Vorname = ifft2 (image array) oes isd. = ifft2 (x)5 hn is wo y Qoplying Ol) iq) way, the user Can display 2D -DIT of th inog+ an onvett the insage ho ro Freq Spati in h no loss Since au is po loss of informotion, DFT is ‘nde ised Lfor developing image processing Qpplica tions .>» “imshow (i); Sos) in = fete Ci) >> figure, > peas Gia > 12 = Qbs (is); >> 13s fftshipt (ie); >» figure >> _Imshow (mat2 gray (18); i >> 142 hoo (1 tabs Ci3)); >> figur ne > inmshow (mat2gray (i4)); >> 15 = ifptshift (3): amiga nifena (it) >> Figure, >> ces >> close all: >>_ 1: imread C'comeroman tif’); 2 Figures: >» Tshow(i); >> Jz imnaiseCi, ‘Salt pepper’, 0.07); oy Figure, chow dds yal>> Imshow(k);