Scribd
Upload
310 views99 pages

Image Processing Full Notes

Image processing notes

Uploaded by

Md Anas
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
310 views99 pages

Image Processing Full Notes

Image processing notes

Uploaded by

Md Anas
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
_ LAST MOMENT TUITIONS For Full Course www.Lastmomenttuitions.com 20. INDEX Explain sampling and quantization with diagram - 5 Explain image file format - 8 What is image processing ? explain digital Image processing and its application - 14 Color models and its types for digital image - 16 > RGB model - 17 > CMYK model - 18 > HSI Model - 20 Explain Chroma-subsampling - 21 What is digital video ? explain CCIR standard - 23 Explain homomorphic filtering - 25 Explain histogram processing with example - 27 Problem on Histogram equalization - 30 . Grey Level transform - 35 . Bit plane slicing - 39 . Problems on Zero point memory operation - 41 . Filters - 43 Region Growing and Region splitting - 49 . Problem on Hough Transform - 53 . Normalized Chain code - 55 . Discontinuity based Segmentation - 57 . Edge linking using Hough Transform - 63 . Explain the first difference make the chain code invariant to rotation - 66 ( update of this answer will also come ) What are Moments ? Explain with example - 67 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39, 40. Explain Polygon Approximation with example - 71 What is unitary matrix - 72 Write 8*8 hadamard transform matrix and its signal flow graph for fast hadamard transform with example - 73 Discrete cosine transform [DCT] - 76 Explain for digital image having salt and pepper noise, median filter is the best filter - 77 Justify/contradict Laplacian is better than gradient for election of edges - 79 Image Enhancement techniques in special domain - 80 Calculate FFT using DIT - 81 Define Euclidean distance ,city block distance, chessboard distance,and connectivity - 83 Distinguish between Image Restoration and Image Enhancement 85 Walsh transform - 86 Justify all Image compression techniques are invertible - 87 Explain Wavelet Transform - 89 Explain fidelity criteria - 90 IGS Sums - 93 Lossy and Lossless compression difference - 94 Run length encoding - 95 Huffman coding - 96 Arithmetic encoding- Arithmetic Encoding Improve gray scale sum - 99 More update of videos and notes will be coming stay tuned Taal , Huous cencecl olerta Into —__ii) Quan treatin Se aThel Tetakie: Folens behind camp |_— Digisieing the coordinate values Ps catled = Sampling. In simple terms, .carnpling aefercsto ' am —eliviching “Hoe Proaye ip equal parts: —_—_ @— 2 Digitizing toe amplotude values fc _catted ov, —_ @uantication. Jre_clividing Imaye ace. to totensity values — >in Fig 0b) 9 onezelimensienal functon te | plotted clepenct'ng on jac jnlonsity value of [sth Image found _aFten ccannige image by —a_scan Line AB. The random variqtiony are | ___dlue_-to_nofee- — — Now we _snapple the fiunet'on cecthy eG ually | — spaced camps along Line AB indicated by | Jer bt fick merki'n bottom of By lt Ce) I -damples_are thowwn e3 crnalt Circles on te £? | in rig at tide oF Ap DG, there ita scale _ |—elileled tate P intervals ranging fom Sface_ —to_wbite Co-7). finchion 1s quantived atcoretiag to Shye _valigy a The _Mgirtel_camples resulting fir boty sompling _« quantitation Gre shown tn — fig ol) _ = starting fim top of Vmage h carrying out above procedu re Dine- by bine produces 2-1 —oligitaltanage. S$ be ice: eal Date | I | oe ee @® — IF _an_ sensing array Js.ca1ed for sampling & quontization| —_The_nos OF sensors in the areay setae _estabitches the limit of saropling_in both alivechions: fig la) | — MMactaty this concept Ae = Figl2) shows ¢ continucns Image projected onto | | plane of -an Abreey £0800, Ag 2)b) Shows image ____af ter sampling quanticatton. pS Wercan clearly observe that tte totencity value of | A partauar pixel 1's rounded off to Weare iin tene!ty ss Value, ss : beth Page No. | @ | naan po | Sple Reader $ > wed to Idemtify tne ctype of fie cS | -Fplate ange Eile Pormats shoe VT ea PO BITMAP Trrage (beng) {J > this is the roost! Commonly used File format, | cHot Supportecl: Yrnage. ranges hem 4-32. bits. a We can use «bmp file in Windows es woll os 20 other Operating systems. e@ TF used In Windows, -bmp File maker vase mar a Nencer Fle called ca Windows. br’ ——| => While using jo other 0.5. Futouing herder | Shuchire ts usec 5 th hee 3 Petes a Heid Sie a Desciption bE Type 2 tharactens BM Heat deeb zal i bP Stee. ri — Sle Size nee | LF Reservedtd 2 Unwed ne) || bFReserved’ = _ Unused _ _ EFOP eB, | ofbet to stort of pleel cata lf © ee bf Types wd 4 7 | srnage 4 Héader$ -_ — IE concict oF 'c enti hy_ mp ax OIE. File. Ip bevadedrnat ie, BM to Ascr1 itd bfStee. 4 wt ice ok BMP File in bytes HPFResened $ Hree ae aeserved, th’ crctiel value clepercs 00. the application tat Creates the forage fOffset’ offiet te. starting aclelves, of the by te where Hoe : bitmap %mage clata can be found. _ - : << tY BITMAPCORE (be) $0 Ia. 7mppzneo Cer) Ady : — Size oF BITMAPINGO is 20 bytes > that of = SBITMAPORE js 28 12 byte. hi tables = BITMAPINEGO @ era Sie. Description fee biStre Ars ghee of tite hreacter C40 yrtes’) _|} —bildidty i 4 bitmap_wi'el+th jn pixels |] bitetgng bitmap helgbt th _pixele - _____|biNoof ColorPlane | 2 | nm of color planes Crowsk bei) |] i No.of Bis | 2 | re. of bits per plael, iwbloh te colon clepth __ brave 45 Lan -oF:Froage. Typical values are 1, 4,8 16 24682 | | bilonpeession. Mod! de | Compresslon rnethod- being. epee _ || br Raw Sieve | A.M The te size oF ran bitmap datas a Luror | - —O.can be given for 81. RGR bitmap | ieifteriaetial Rea | 4 | | Hertker tet ftuohutiva ob. tive fac ge 4 | Bi Verbal Res Nertical- Fesolubiva of tre. Joage —__ © LiNo.Color.Palette! 4 | No- of colon in the color palette | # or to deFeutte: to. 2% sat 4 bINOWIMP. Color 4 no. oF lanporkint colors used, or _ 0 wher every color Fe. jnmportant | _____ as ne BLIMARCORE ‘hence contains 2 st Fete oF tne BrTMAPINEO header, 9s) | Cotor olor “Tablet _ — color able Gpaletted occu in the BMP image | File_olirecHyaFte> MP file header = — Color abe can he of #0 a Rcequas | | enrRree Fel size | Desciphion feta Sic Descripttor _ ——_rrgbBtue Bhat color pete | hdr 1 brreen Color Vatae 1g)Blue 1 Blue Color Vale | rgb Rel Pet Color Value _xgbGren \__ Green Color Value }— a raeanaa rust be cern: lt Pe Coter. Matee | RGBTRIPLE 7 a | Pirel Deka 8 stiekepioe ny =) For spiel. its we colcabibe. raw. ht Lob Soap ovale oF Height C Frew. Stce_can be ig iven.aa. os aHlelghtn bIHEBMY Pay Sig 2 LiRawsie | wa Baten, . be Raw Size Le a Bae ring bipts: perntous! BPR = effE (chiar tinsmesertpafogig—— | Wes @ || for Abie there are tivo rartanh cleveloped_by Trtel__ Tagged..Trage Lele Forrnat ( beth) ee >? TIPE is Flexible, adapthte File Format for handling Images A dela within a single Be, by Inctucting header clags_cleFining Ihe Frage’: geometny —— 2 Tags are nothing but sice, resolution , nar Ete oF Nenetge hoor Woo bytes o | Intel] Motosola Byte Oreler a 2 i Version Nam bor * 4 Pointer to First image Bile aliechiny ——& ) - ss T dmose File Directory TIFF HEADER SRocTURE Byte: Orcler = a - ee 9 Here we specifically took after sequerte [orcler ___which_bytes are_storeel. Pa __& Morrolo ab | MHD [ =tag xb | X44). SData Type | x46 I Length | ota Vatu of Pct X4E ~7ay Tb X40 Data Type xr) Leng to KHlk Value offset X420 No.of Bytes Bary ui Divedtory Baby | Pit pS ele pip Divectory Labey 5 —\—— next LD { JY PEG File Format _ | — > SPEG Fe 0_acronyes ‘for oink Prati — Experts Group > The SIPEG comprescion etlgositomn feat it's best eR pho-togrephs paintings of gealivtte stenes I wv emooth variatons of tone: f colors. PS TPEC File clo not heave Formal berclers, bite Fg jpeghead0 Fy? jpegheadl) detcwns relewrt Information fom tte files stort of trome Segment TThic_ te called Heacter For consietency with other Ge Formacts. —__ Senage oO 2 JPEG start of | Jeong LS0T)_creicker 2 i 2 Lanqge width Io_pixets 2 | Image Hel bt Sn_pixets__ 6 | No.of components (t= Greytcale, 3 = REE) | (___Honzontal [Vertial_sanpling factor fon cemponent bo | a —+ | Sempling Factor for component 2. CREB) JF Sampling taetor for Component 2 GE AGES Page No[ | Co Marks) [what te IP? €ypledn Digitet TP € i ts applicebtons IMAGE PROCESSING® : | _____ Zt _fe_e_process _gubere— an_Imay cepkired based _tpoo th require 08 ,_lmag. he processed. Digktat_ Image Processing 8 _|Lau Arte, An Franape. may. be _olefined 41 QA 21> r | Fenction F C4) where 2% £4 etre spatPat | Coordinate, amplitude of fart ciny ofe ____ |oF coordinates (4, y) fe catted Potendity of | Hoe Pnnage at dhat porate | her ay Lb nkencity vatues of F_@re_ Uiscrete GQuantitier, we calt the 2 Ravage ee Dig Ftat_ Favage. “The _Digrted_Laneige ___ — | Processing refers Ee processing obigiteel_mqge by mean. of _a_ctégitel_conyrcter. = = — oe Steps in Digitet res [skip if asked foc Meots) a _ le_Image Acquisition i _achlon_of Retrevirg an. an == | Fmage from _chare. source, usually 2 blo based | ource, 60 it be. A Shrug whatever | __precescing need _1_occur aftervarets._ | Ze Image Enhancement? process of manipula ting sjrage _—__ 0 that the result ts more stteble i Phe 7 | origtoal for a_specthe application £4. dd > get 7 ___brightersarage.__ _ _ Date | _|| —_parb_or_ object. ae + Image Restoration? cleale _extth improving the ____ —Ctppearane of _an_imege. Chale forage enhancement, _ which '< subjective, Image restoration Peo bjective. Bestorato technigues are based _on_mecthemeticat _ models _of Frage Oleg areca ton 4+ Color Image Processing & Ad Gree: Hat hes gained in lig heel Foneges over Internet Vario _mettrods_ceseel here cre color complements, color ~slictiag, Calor corcechionth. S. Wavelet Processing’ llavelets ave founeation for = Fenpor tance becouse of -S1gor' fico wereage 57 ce of meray yy Jancyges In wares clegrees of _wesolution - It tx See yeod for Image ata compression & pyramictat Meppresenteton of Poy jes b. Compression & techniques for reducing storage requived. -to_rave_an_'mage OF Aoe_beret bouscheorct Ho repiloce _-to_treinsmi't fb. 7: Merphotegicat Processing & cleats corte trole to_extroct __ image cormponenta_ Het ave _cosePul Fo_represertetion | Lk clesertpto of foe _oheepe- | 3: Seyrmentation 2 park? G00 cn eenage Jato fb Conctiteent- Ds Represeote ter L4 Desiriptions Pepreseating Fc Pereage shag —— Bounclay vep resentation or Fegiornil rep resera-tet tren _ Descriptive deat with toutes Prat result in Lome _quart'tative Job uttos of /'rterese ject Peroynition 3 Process of ex se igning Jabel tD_ar object pened 00 tts deseo Applications @ 1) Trrage-sht-peatn 1 > =c0eominy, bluse! A Bestoratlen bh er pe calew _dekectirrg edlg<4, image etreval, Prrcue recogni oa 2) Mediceet Piel - - _ » PET Scan, 1: ey inal, —— gamma rag imagiy Medical CT, OV imaging 2). Amazing 7 — +0 cletect jnbashucture Alarrayjes eetised by ap estrtoguake — A) Machine | Febot Victor oo — $) Hurdle Detection ee 6) Pattern Reegnition a _ VY) Video Treeuig oe — _ Color Models for TF igteet — _ (or Maks) There aire 3_chi leven cole _mecctels_Hacctcire_tused_ __| accoreling tothe capplicertions + 7) RGB Modet ———EEEEE i) CMYK. Model ee WHS Medel — “clay K_Mocet 3 — —2_consict of eet en Magenta, | Yetta & Keirboo, hence lemon ov aetna __cetor Moclele —=>_Useel_t in igit_pikny & phategpapy ujheres —— | RGB Fs used Ter comeras << Comptet —_ ( \ Rg hs} wt B Date R68 Moctels oe _ a —_ toidaty _tued jo color TU seth, R -Reo!, B-Riue, e ~b¥ eer 2 RER_ar pamion Colox_hence_ PShie_ic¢erllect Pomnory Modet |__ > Operation performed ysiry ROB rundel are easy =) Ib Ss_hbased on Cartecfan_Co-ord?nates. System. _____ —5_lAlitin dis model oe_arequire 8 -bits of yedl 7S -bila_of | green _£_g-bib for blue. Overall ROR model Yo 24 bit nvodlet. 7 = =, a — > Ny picat_woay oes Keprasenting- Roa fmage 7 ex) tuhere 20°F y _ts_stine physical sixe of yage dsb tom |_____planes 8, G,I&8 So_we_can_Say Hot _an_ RGR mee. 2 mage is_q_sstack of 3 planes of size 21% YY — |} = Not vey yeh fac Pmage processing nts — - oF RGB csnnot be treated i artependentty ——— — B, 7 BS eo | — EE 71 2 ——— pragyentay <1 fe 7 1 _ 7 fg a — ~ | O i foo ee ae a ReV(1,0,09 = ———————EE = oS Page No.| cHs1_ Model 8 = - | ——> used by lu can _eye— = — > Aue _eye can determine color Using +O%< ________ model where sH~ Hue Uhilon i cte Fined cts _ov _A8nt or _cimply— color ov shade which ___ dominates Hae} Maye. _ = ~_Satueretion , che a mount of white Colow present ta sae put colon | Hue + — pS | > 1 Intensity, brightinen of Senewgye |_ > it _clecaples_lotencity fiom Color foformation |_—> Fdeeit fon devetopiny Proaye processing. algon toms _—_____ 2 Considen_an -% of a _bucteet full of red _colon = color red Carresponcu tto_hue & Sctu ection js maxim. When ur add Li'ttle white colon inthe bucket of reel, wed ters tq. ge2ta. : |____$0_bue _betanes Magenta € saturretion red ees bo OH SE we adda bucket of be white t_ — ib ber real tucns +o tohite-_ _ | > Consteler_cno+nen fy. Where there tsa budeet — | oF Reet colon, JF ewe aclol Mittle Rial colon to $k then Red stuens to clare red _€ 4 Pus _ i |____Keep 00 adding black olor Io the bucket _ oa | nen rest Color till kuea_into blacle, > Bo ue cin cay that dark od 6 Magonty ore _ ati Ffement saturations of red cotor, Date[__| white Rea Black. pie Sabinetion Rect Nfettew aaa ine © ie BSG 260-0 iF &B>G © cog | = L@-O+ce-9) = Lcr-wo% ce) (omen) ; = . ee eee wee S : R4+GtB l a : I a Page No. luminance, Cinmountof_2Pergy ‘rece! ued _ fam source OV-= Chrominance_lchroma_, combined OF Ate & satutet'en ____victeo. processing jo Tl cet. Sie. huma tuces cued to emncienb = jee. “Black -o-~wohite “Ty sets > huni ioe eed “toad {just _bs'yhtney isthe _| Block G vohite Tu sets oe — systens 2 UV color_modet_is rcunty cused tot 19, Ptect Chrom_is ued to add color intormetion - tn SE leick=n white TU sets. = _ — 2 MOV model is usecl in chroma Subsonpling — UV 'mecet_ can _be_denivec!_ from RB Hode wohich I s Follocacs 7 — = - = - = Y=o-493(R8-Y) _ Y=o-299R+oss7gG tose Y= 9-8771(R-¥) — Se Senna - ® Date a . lain Chroma © Syb sam, = Chinma_cubsanipliag te “aie Sob compression. ___ = Lt it _elifferent Bom aspatial L temporal compression WH by mock video com cston algorithms dn clokeck Crease oF tte ‘mage’ in_a frame, or adiocs a no, oF Frames _evhere ? OF Fill herieosil) Reawtutien ah Vertiat Reoluton Ja he Feentet resolution! al) Verticat mesolate g Hovicentat ve solutt Rui Vertical resolu Hon, Chroma lubcamnpliag. hes 4. vartanint | jt) 42424 4» No subcampling fi): 4.3 B22, 2? ech horicontal scanline has. 2 Chima Values cent for every Luma -valtes Wi) +2121 > Jf emmma velue seo er aery + % Lume vetives Seu) 48.230 9 for every “4 Lume verhees rere evokidel be 2 Be Brat chroma cenpnent <0 fee an fon which will not .cromt "olen image Fours cassumre 2 valee Fir per Stantine by Oy /olr Image oli ABCD \Ato. B+ G42 b43 s| |S 6 1) EF GH Eta Fas Gre Hat) S ISIS U0 4 [Avo RA B42) Raa \e*4 | was 5 Dt aes [ato Ane aaa) Ata ih B+ oss, +6 +6 o+7 lato ari | at2\ ets ‘le 0-7 Black to white A-H dome colors ercucr jatel ets st (Arg A4S BAC (R47 eee eee 4 | | \ z * | i @) pate| || | || tAlhat Pe _eligittet Video! Explat CCIR standard Digiter Nedeo® Meee ea sy | > Digiteu _alteteo is fenbieta esl ahetag tng Miscat | tages in the fxm oPsadsatedaligitalecteda. 2 Digital _vitleo comp rises’ oF a seriesaf orthogonal bitmap oligtial tmages elixplayed:.'o repid seecession at_a_conctent pate, 2 | = Digital video cen be copied xv no_degradetion ta, quality Mo matter tow roany. generations of a clityttet ssourte fr copied) j/k sill till he as Clear cia Hho ontylnay Fink generction of He lig tat Potage. ) Advantages § é 5 _ i) PRarctom Direct Acces > good for non-Linean suldeo ed pH fi) noe. problem for pepectedl gecoreing i) no need for blanking & synz pulie Consubsive Comm hee 1 Fon introaoa Rade (ons: 7 _ Ab hoy wvertour verstor Core ~ 6018 |__ "et Ups Tolerlased amin, sn onc Belel only has boll np tsk = | Mertfont wevole tion, [sp cee dette etait sppaninnty 165 hp = _ttseel__for broad casting Page No. 6 jo Ln System BR, In each Bowne there ewe. Chars stem A 3 ee mi | > a frame ctotett “pletun. Fame aute te — Pictures cesplayed In one_second. But each me_is actualty scanned _fevice (oferdecsicg—) Odd ever Lines, : Bach scan te known as a sLinaa, ade feta rake | he Hutte tie. Barre Pate, e- 4.05 Line, 60 Line rate. ps B05 times Bare Frequency ive. wt bine Frequency = 405 K20 slog He 2 dlere video band vurelty ty. G04 Hz, 2 the channel bane eolel Hy iy’ £ MMe. 2 Used tn Color = wed AS soupye)_mne ister | COIR. Syslers be _ He otf seney 62S Lines. per foie «| ® w+ line Brey uenty = 625426 = (S620 He 2 Video Band width 1) 5-0 MHz Fourier Transform] fOotyd] x Rourier Tae ep) tale apply Loy both J tales be Flay) = a flay)» Lovbleesp) |___——» ov applying Fourier hans forma Ht gek- | Ef Loup) = =F [tp ae za Alen ol bee. lew) EF; TE lev)- ee A Where yp fi law) Fe nee "fourier pancho 1 06 hp Ebb? Lp Oty aeapectioely. PageNo, | ce "Now we cen copplay shomemorphic filkee ca | Suv) = Z(uyv), Hluv) s a — Slav) = Ey dav) Hla, vd +R lav avd os Bitered Image in spatial clorraln is F[Sta,yJ=F Cex tava) J+ F [lav lewd] | e he. | af 4 | Sly) = 2 (x,y) +o Guy) | oe ZCuyyd wa Ferpoee/ ay taking netoral Log oF Up image, Bue Mevene the process by taking | exponential of Filtered result thorns olp Jarry e | gery = eo | bee Ginp=< Ble nye 0 (x,y) , temy) ony) Leo whre, . ony) Lo (Gy) = & J 7 ey om Cauyd = The Filtering approach Jost derived. Cn be Suromer rit cect og? resort enf er Hfoos apt ip—— lp fmoye = image ——— Date| | | | Explain -Mictogram Processing ust Clomasts) 2 shistogrtins of cr_ctigitat image wlth Sotensity Levels in Hor mange [0, L1J is 4 cliscrete functen bled where ay is toe K'? Intensity value & Nk &s_No of pixels tn fanege volth intensity Sk. > fs tog ram ic_a_globat_descerption of Image | Here we plot vatues of tmaye ort x GY anis, 2 On A-ax!s we plot Te vatue Conteris Nk vatiey. = There are 2 clifferent Histogram proceacing techni quead_ 3)_Linear stretching ii) cfticto grams Equatlcs on _ MW ble began Spect# cation, Linear /-Hictogrem Gbretehing $ Sere main_aim se to convert be Joput- image. j0t0 the high _conskrent Image. | for a's wwe _mcake_use of slope which 1's Sax - Smin Tox — Tin —Slepe | here, Smax.8_ Mare. intens ty /o_olp _ Senin? Min. labensi'ty | olp Vimar Max Intensity in fp EesEEsEIDEsESDEINIES Vp 8 Min_intens'ty in_tlp | 5 Now fines etretehing will be caleulatest os (SRS RON (ra Sonic) tb Sen Slope Germain’) + Soin Page No| | for 0) | Grey Leveled OK ott © [too] ao bo reinge Co, 7] | Soe SS SO —Cmox = Tmin =O _ |_luobite electing Base, obec for grestet apeezern Tp cl keg Won-cevD Wie veetue “ || Now, _ LN Cr od0 _ —_ ar ° = je. Sage usen't be _ _witn Lines _sbretehiag, we get a_mork _ eee 4 image buck ee chet ee can noe _ | 5 for ~Hate equatirertion orethocl is twee! + > In _ordler to_get Rot hictagram, ase 7 conoliHony roust be co-Hebied. |B ros must be motootorialy in intreaging 2 a single “valued. ———— ) “Range oh Jokensity value shoulel_rernge befvoeea ______ oO, bi] je. O< Tery Next we aateutebe Watue oF Geter Salt yo 3 seren our off the vate of Se (bet) av eeUGY IT as=> Date| |-Brblem on He, teyram —Sualiex Hon Piret Vatwe [70 [100 [eo 80 | bo{4¢o]s [2 [Gray Level | o |] t | eulal als 6 mi Hn | CE (Se)) Sell) ows | pees 0425 | aegis | Os { 3.675 ozs | cow | OBIS) bles Oo SW | 6-827 ot | 6969 NY =Z00 Mow consider 1, 2° 6 Lak column | Od teeny Level Pixel (ne) | { ver Dato sistoyrein Specification * Mere we Gra ~equctlice the bots histoy reins. ——* then compare eq ueliceel values STEN clo BeLene reapping ty. 1 1 | 2 | a, | z | co \« A Mk | 19° | 1073 | asv | Sse | Bra | aac | ier] a1 & be) oO] 4 a2\|s 4 Ss 6 4 DM. |o rt | ° | sal ‘| Shs G14 777 Histograms «4S “Histograms By we | me | PRFLeDE SKL-) ROME | poe | CDF | Se} ° 190 | wrig/ond | 1-23] Yio ° 8 0 1 1923] 0-2} 0444 3.0313 lo | o Jo ° 2 | gsp fo ztfobs| +ss]/o}o | o | © ° 3 ose | oth fos] 67] fois forsfows | hos + 329 |B Jo. R9 623] 6 819 lore baS] oar S eas Jobo. 95/665 [7 par] 0-3 fobs aise 6 122 [00 | 098/686 | 7 laa }or lows] 5. OF q 2) loo] lo] tro] Tou lous] eo] 7 n= 4096 - ne foe _ mo nk St. ce ne = St ° 30 =e £ ° —_ I lors 4 3 ° —~o 2 B50 a 9 3 sb 1 msg 4 64 | 328 j= + sy Ss 24s ig 122 =—aS ae }22 js 6 B19 >& =) a\ u ee a RO ° ° 8° PageNo| | || Now perham revere woppieg aE, = Gonnpare— Brotha Reaitiecticect—teistog rang OF a a Se RH — J tresSe bold) ne Sat ettea io A) OOo a 190 es Bee t | |@ lo ARH 4 | |p |- eC ( Ja} a} 4 Cray Level ransforms é JA ar = fahancing an imay — ~The hantfecmetien £2 Ki 9 ‘S=EOrya ae where a3 pixel: of sp “$2 pikels of lp A eae 72 tanct'ee that dremfonne chch Se eter 1) Logern' thai i)» Power Las miei e Lt rmahion 8 has tuo ty. pea 8 ef 1) taentity 6 : a ests nae DMNNeAco-not 4 —o. fds Hr rey i) Newall Ns ae aa op we hs acini pe ES Ag stew ok —2 tei Hon is srepresented ¢ S= b=) 0%.) = ee tubers Lot range. of Tagen. — 2). Logan'thnsic TrancFormation s. sesh vice . generat Farm of- hens Formation At ( 9= ¢ fog CI+¥) where C3 4 constant which we add tote =9 tteahah. sha typen se iN iets 1) Log Transitisa : Ww) nerve it1on, clask pixela 1 tg) Date| | Power Lausg . : ——- 3 Hes wo bon tHany 2 on? power ft ook —2_expreased oy} Ss cat = a wher hb is. called As Gomme clue Fo which Hata. ctremsformation: i cette ey Gamma Trent foc mation > Vestation tn vere enhancement oF image —_ 7 Conthat Stretching * > @ process nat exparcls range oF Sotenuitg Levels In image so that TE cpans ett totensity ©. tn Tecercting meeliim { cliplay olevice Here clarker friaga are caphired clue to poor Maminane oF hight — Gaph ty gin by Lig. ps clvtded Jato & regivny | ‘0. 00 2 25) a a ° ALC ; : Oly, 82) ple) Cle Mss G4 =$2 — a tue observe Pparperty ab eden tity “han Poem | 6 wy ea, 20 Sabo) __ SS O12 “Thresholding # - > provides_an_easy Wa On the heat of — ab ferent ingkenci'ttes or —_ —Colore_ip foregreund & beets ground tia te => Seg cnenetartines ts cleterminey by @ “ingle = -perrarmeter_ finoesn 23. Intensity Ares hotel. | — > Tho single pay each plrel value is —_ Compared with threshold, TF pixel intensity | @ i ic higher ~than threshold, then pirel ty set to ccay white in Olpp ele black Feat, Hhreskalding can be represen tect = -s2[0 y bet _ A | tuhere 9 threshold ~ Plane Stein => pixels etre cligitat numbers comporec! oF bits For €, Intensity oF e6th pixel in 9 266-level —grey scale __ image i's composed oF P-bita. 2 Here Instead of highlighting Intensity Level_aanges, Gy u we high tight dhe combution mace cto total Image appearance by «pecihe bt. jee. We observe toe Changes jhe Jrvege bit hy bre One Bit Bit plemeT - Byte P| Cmsey - Bie plane O a cLsR) a TS 1T-PL9NE Representation of 8-Rt Triage FP Te_ebsetve charge bt -by- bit un Sirat convert _ the fobers ity veileies '070_1'&< binary 2ocu'vatent _ Tn _we_slice sthe_biinary image hom Mist ty LSk. Jres_ we form _ob'Flerenk Jooages For wach Lt bef intent ‘ey. oy ie - — [ssolvory ry ria] = >] 001] 00 | 100 6] 3|3 [ego jou [ou Ms Middte Bik Ls o/o|t ofrfo}] |i to ain ofe|i] — foft | oo ff vf oft. PageNo.{ | I |_ Tntensity Levet_Sticiog-8 _ “Highligixting_e:_spec! ie_menge _of_lnkensities _ No ein inage—of bea ts_of fkereaks Applicertiba_lnclude. —_____ Lenhancing Feaketres such a.1_asses of _werter in | cateltite imagey € enbancing Pteuve io Xaray trates “Te_procen, of ben cotledl tn tem ty~ level Lclicley, ca be_lanplensented! 10 several Uses | but mat are van'atiooe of t00 bowie theres» ———, Ore ep precich te to_cti'cpplay JO_one_valse — | Csay white) cte_-bt values tn the _pange— of Woterest || ein anoter (say black) all other jatens hey. This. cranchormatien js clown 10 Fig la, prodette —_—_———— |e, br Bf janege- — Tie second approach, heused. on creas torortfin Ja fy Cb) tenightens Cor clewkenc) cleyirel 2 ge of traterae Hy. but Leavey all Ofpen Intercity ténchang cd. ——— tv | -t-+ _ v a 1 E E | 4 _ ZA Lf if a Ss jen & Ter) | - | = L. ——— 7 | — be j-_-_| . —i 1 . . Liste 1 1 ~ 1} at 1 1 ay a oO A B iW - A tt Fig lo By CBS. || higntignts intentity —chightighta ange [0,12] — : ind preservea ll 9 ther intenc ity levels. |_remje LA,8Y tedvteg CU Other _jrtentities bo. a Lower fever Date | Pmrblems on Zero Point Memory Operectior! = Silt 4 _ No O19 bY O10 — vol tor] Goo |] Vay _ too J ory { PSO | Boy a pro fron fii | uo |. [_— ma ice ; Be Bit es = = - tensity Level Slicing with Beckgevund s Ronge [3, 5] Here we ae palace oe_lo kerncity Jp _rannge £3,.6. —) cut __ | Phe — back grouret seaserncity—s'-e+_by 7. a 7 Page No. giteut_ Meqcctive MNGi tat eguttue oF aay Trresholcting ® with £= Plodmmexpeptete Vere : —— _ ee | Li 5 SSR _ a = _| —_tir - _ Ss : ot w<3 a — Sat > To8 wes ee a|oO}]o]fo _ - o|#7| ° | 7 - 7 7] o[°? ° a a p ‘o} T[ 7 T Filters @ There cre typeof filters — A) Low Pecs filter (Lep) | Smoothniny Filter J Aigh Pass Filter CHPED | Sharpring Filter — In Low pass -Frlter, It blacks the high frequemey pixelc © attows foc intensly vates pefrequeny pixels 250! asp 280|e so]? S19 | 240} 270/25 % 2xD| AO | 240 | 280 [eFolleso 2Fo|s0| oP \O | yo | yo | 10 wo} io | io fis yo | yo [19 | wlio] i fro fio yo} yo fie {ref iof iw] wlio 90 | 90 | 90 | ¥0 | 30] 90 | 90 | S0 W19 | W490} Ho] MO | 119] 10 [10 | IF ogo | 260] Wo | 20} 250] 50 | 250] 250 2f0} r50| 20 | 260] rfo| ww | 29] wo oso] 20) aso lero [efolare asolaso Date CIT — Ey = (e109 0, |bOwW Pass Median Lilte 2 low pass Qvertging Alter, }f He famage oF Size we get_a blur Image, | 2 TS ove % ome th Lue tigi big gem ~ Low medfon Filter |>_TF. the ieege Conc ict of salt oe Pepper 20!se, Hea we || ——-tece mechan FP i fen 2 Steps to tue medias filer’ ne FB) frecune 2x2 Empty sk 1) Place pousk of fanege. — Ati) Arerenage. values €rttem la Ascenchiay 1 Beserrcting - WW ehvese a median Bem Hroc6e 9 values YI Ke place rrectian yahve 1n_Ouckpet saage SN ) Cootinue process for re bole _jpmege 1,19 VO 410, SF wD) Mea ve yo & oe — Fj Meslian 19 - Oyo a lo “Sfie} G/[z50 | > 55 je [aso |B] 50 [Ceo] azo) to | 10 fro wo | 10 19 | 10 fo | 10 J to yo | 1 fro | yo] to 190 | 1 [90] golasol 20 | 250 | 250 High Pass Filter § Jhera etre applied or tied 0 —> After high pass fis On ilp image swe get mor _enberred. edges J Abe olp imege > staoderel mask For High Peas filter ts the centered Value — Tn steneterel HPP meat tg ue Gant other cure —ve - — Here we clucays get cretion of meskes Sto - “hy. - [ofr Hoy ve His -o- re 9 [so [v0 mweso-w! ig | \g | 10/10 57270 syerte 19 |S . =o “ > 4270 1 Wo joo \oo yoo ole = Date to Olp Aamage wwe get veutues ts +270, 270 whith —_ }—_etre_ oot orange for she imager = To. sdve thc problem Shere are two coletionss 1) Replace —ve vatues_do_O a = Gb tls observed thar at the edges values oF ple are out oF range. 10 avoid Sole we tse G sealing factor which hetps to get tee _verhie_vot'ttain range. te _Luill cise Scaling Factor~ ero the mask & boty col a new mut) doe _eutcpur Foo -_ oe} cfo]leo]} ele fe jo_ el elalic)eicle @ 7 7 of elelefele {elo — _@ ol efelo| ofole |° _ 7 | 0] 20 {20 | eo | ze | 20/20 | 20 _ 7 : —fefefefe}efesese | 7 ; ofejo o]e;e}eye | _ : . colo lofe fo jo jo |? 7 ore te ee ange — 3 Here are PageNo.| | thigh Boost Filtering: HPF hetps to get ooly Forey rpund_inege esol, p_________ | beet if we want foregwound-de coetl cy back gepun ___ image tw olp this technique valll_ not _wosts, tp = |___@vertome this we _u HBF, | Here we pass high frequency pixels along with ________ Lov Frequency pixels. a — | —> we know tert, __ oo High Pass = Ontgioct = hous Pass ________ | fo-Bind HEF woe use mils pliicattye Feetor A High Boost = (A) Originat ~ how Pass __ ba we get —_——— High Boast = Hig! Pass : _ |e. chhigh Boask = Original Lowe Posy | To get AiR E ee vette of ~A_?s_celedeyyc pber aa i ———EEE If g Azlsom high freq uenty etre added oto > | original imeye. HRE = DDO ginal a Oviyinat = boi fs i 3 Thic_bechri'gque is called Unsherp Pleskiag = = => Mast j4 given eu - - a { x] -) far la ae sere in shts meuk QU § Valued cre —ve w berea: | cen ben verte fsbo tobe found K= 9A) =I — oe @ Date ee : BJ tt hast typort _| Singh dinkage 2 Kegon Cronsing ch Bayon = Sabregs 2 ead Fapton Region. Growing § : |} Here -we— join te piret -by_comp. aning the — —hetghtourhood veetues 0 Find whe hether they ~they setirh, the threshold vatue Po Single Linkage WY Centervia. Linkage MN-Hy brid Linkage Lourhood values? Wwe We 4 = connected method: —> Here for Compaping [oe re La 1 [a 72] | ost sel | aw] Re . joj las fog stro : —| fog jo? ez | an | Centaid Linkege é | > Here sue fut Calta late mean value oF region Lo fe | hen compare_tt tutta news piret Dtk her 2 cases. & PageNo. | © when _pisel_valie ti Clore to mean vatueok |. Neighboucing segmmenk. | New pirel + to“ u ee [ou a] ee gu 0 . sae [oe he : i —- M= 1027 — a Date Gn) en oe | © _Pixet vate oo: not ft In _2ny region _ $s af fe Pa o nou a a Mm = 30 may Proll . 7 New pire = TS ~ create bow region [75] maT Hy beiel Cobage! —3/t tag maximum Crkelyhood dechnivue ie 7 here tue frst Consider fevo Pikel_vataes we Find out nefghbounry ish of portical pixel, bi then Mind! Gee caeatinn piven per? bot Cyt TF we Find More -han co% Commen Clement — Hen maker rt Fearn. az 232,31, 76, 33,3,32, 31,92 4 —_ bi yp 32, 30°, 32,3), 34, 30,30, a/F vs hore than £9 6 -elemenh are This _ne__rnate Tey hoa Ri PageNo.| @ Region Splitting & M 4 e ¥& ferginy ? = considera whole Inge L cielde Stet image | ve) 4 parh Se ppreshold_Poeyn't lcotis fy = _Contiaue qbove_ctep argh 0 farther Splitting’ tu paces bler— —— = Per foren 163.400 merging by compariay _hetghsourhoad- ppitets 4 merg LE Shrahold, 21H hres | — Continue 9 bove the int'l _o_ Further _ ft mrecging possible. _ de . e tho Feqtony qo t =\0° — 7 ; tee) 2 (0,1) 9 Ao, fe c || jf eso, tf mzo,¢_ _| fer _(-1, -2) _t=m- | fF ¢=0 if _m=zo — y &) Page No. See “The gia _point_are —cvllinern ay all three _ _ Aine _tnterseee ab (3,9) et |_ we cubshiue m=2 C=) ti &m | - A —yem x+e = - _ : oe = 3xty _ a ee By? oF Cree — _ a oe as a Ge @) Dato| || NMormaléred Chain Code 8 -Mormalized Chain Code —- 5- connected system SiunnionenanNEnenNEene = 7 - i : Ss tham (nde; OO BS aoa, = 7 ___with oreler = 8 — | for cateulerting Normabieed Chain Codes || 7) calalate chalp Code | 1) treat chain code np etreelar Lshlon HW) catuslate ciredom Rest oltfrerence iv) caltulate shape oes PageNo.| | ||_chain code Cec} = 03032211 _CED = 3193/0303 |_ NCES = 0303 3133 vor, Changi ny starting pr ccf 33221010 GD + 03033133 NCED ! 03023) 33 Neen 2 93023133 : Le! /0033 212 if — 1 ‘T C£b? 4o3033I3 | = Tin_cliseontin sb rae approach, he park OF: Frage te based on some abrupt Changes !0 The. ff soihan sakens ity of core _ — | —» Here we mainly terest fa Ielewitfrertion ok _ Lsolatecl points, ines k edges tn trnage. 2 for An edge fea boundary bet” 400 regiong baning —___oliictince_latener'ty Level. _* TE ic very weet tn cadet oF alitconinally in = an In fe in tmige Pane Ae clark = coh) _ vice-varse, < rh (o) intensity profile Gb) Venter derivative ore w@? order Glerivative PageNo| | ED : | Pisk-Order Derivative i atime by 2 th sill? a a ‘ : Th = gyal C2) af gap a Ta Sheng took 7 Fs magninde 06 (6) _ __ @ EV Gi + Gy" sul oe beg Ee = Plath) ~ F009) < F=gy _ oe FLY, yr) Fo4) ee a eae marew - t o | =! _ - - a NTE Sant Efe} pal | fy Netious masks used for Boye Detection ares, _ D) Roberds Mack — Fal ze-z9/ ok ee | _wvlFl= t _ ee = = 2) Prewitt's “Operator feeaueteementee oe fa a [zr +28 +z) — Gee - @ 7 fy = eee ee ees | - y 8 Es ce - | “A fol]! ee _ = -fo |) ; D Sobel Mesk & Fas |\(z1+24e+ Zo) — (41 +972.4751 -\ [-2]-3 o}fojo vilo2qy | 7 “1] O45 -2/o9 | 2 a - lol, a 7 oe patel [| | Edge _Lfolsing using -Hough —“Ticans Force —_____ | — Ore powerfil global metirod lor cletecting_edlges is ____ Uled —Nouge _Travae Ferm. __ ee i Consider co “pont C4, 4), “A line pascing teypugyh point — Gy 4% ACY. plane_can be swxitten_ja_slope=iotercept Foran aa —————— - yh moxie m—sloyee, C= 10 tercept- | 5 using above ~e & varying tne vetlues of © ©, am inBol te no. ¢ oF Lines passing the polot_Caut, yyid-can be daawn y : | > The above 67" crn also _be _eva'tten 10 Eerars of Slope + ni_& intercept C cut a me ye 2 Lt can be plotted i9_na-c_plene (Hough spaced gives singe lire 10 me plaoe fore tired poiat (Xi, yé) |____ where At &¢ ween a — CHa FYF — +m | > similenty fox another pein. £0 544) there will be |___ another Une fn mic-plane, _ | 3 “koth point (01 s4yOoS_ Coy yi coil exist at the boterecttos_| | __lntensection of Hye lines ct mo & 6 in moc plan | a_Une_in_2ty-plane with 1 clope mf. latercept ¢ passes _ chorough_botn_cthe_pointa+ ft + Thus___| Page Nol | MI the point comerined on came line josey plane ——____ muct _heve_conresponeliag Lines ta _paranreker — pace Shak jsrtersects Gly!) ee | > For keeping Arack of alt poinh pasidny Mrnagh chaight | Lines in an Image, Mme parametne spoce ts -cubelivicteel tate _actturetators cells a3 chown? coos | == —— | wwhere (amin, man) fe Comin, Crix) are cx pecte cl | eemnges of parameter values For slopes mmc hrkntept e.__ I Cel_witin coordinates 1,j has oO _clecurrulator Velue ACL) Each actunula tor represents acpecih'e auage oF “phe | —)_=% nitally atl A¢¢linajator ell ar set cto O- _ —>_ for “Every —poiet Cae, Yr)-i0 fie tenage Plane, —_____ ___-for each of the cttowed subthuiclon galue DE ale pe he to_aceumulerto Cell 2s Hoe _Vertue_of nearest Pt Cie Worked out, & wounded off +0 Viicwved Value, Date | — iF a choke of clope mp for point Cn, YE wesulta in esoluHon ob jotercept Bq then eiceermuletor ett Alp ,g) is Tnltermen tec At sre end, value oF 0: Qin cccumulector cell AU JS Roeticates she Q_ points tying on othe line, oe PY EON EK] tohere mt_ EG ore range slope € tpkercept- _|| covered by accermulator colt AC, }) = — bq, let_there are_toree points (0,19, C1,49, ¢=1,523 a |} —Us.ing hougn theineFesan_voe che cle iP these Point are Collinery we (0, CS O+ 1 =} a ne 7-m=0 (rn, c.) =C0,1) || fer 4) es met renee - 7,092) PageNo.| | | AN the tree Lines intersect at point (31% oe Y= aa +1 is sthe Zquation of Line. | Explaia: The first chifPerente makes the chein code Pavartent to rotation = 7 | The 4 drectiond chen Code_con version ie shown ey! 1 0 z | ‘ale fake asterrtiny points bet ctetting point be _p, sAle . |_wnove in _Clockwice cretion quer ume - I 7 : coe E Me _chate Code will bes | 100003221232 Date Shape numer ts clefined as toe smnetlest enogoitude oF circular Rat dlfference Corp y. IAle hene need to Find out Pit J oll FRerence € circulen Flat difference 7 Chein tore 1 | i] ofolo]o]als | ae Fine diferee|! 9 | 3 | o| 0] © Blo}s = 9 |e] a3] ole] e/ale eis eofofo]sfolsfjafi fi fafa fo [a | ! Ahat_cire Nloments { ~fx plein eatth ~eo, || > Moments _are__ctused_sto_re present object w+ & ovientertion | —> With the hetp of moment We com Find the shape € | Sire oF te Imege Se _ > for image of cize 9x0 He moment ot C1,j 9° | —_poclHon icf ten Fly) ix_given by - = oli,} ae cecly f£lyy) #6 14. mto,0d = ee fey) 4} mod = a Ae Py a J mbon) SA y Feuy>_ Jo De St ay, Fay), PageNo.| | ‘by, a = a keep any one rome constr Fouy) > keeping % constont = (S442 +0)4 (2¢0+34I) HV HR4140) *+U4+34042 | (0,19 = ey ta, oD) —— = (34242 4D42(It0434254 (24341 +0) + ¢lot+it+o+4+25 = BSALetelei2 —mlo,2) = = y? Feary) = = 1 Ce 24e4 + 47404342) + 32434140) ¥ 16(6 414042 HF StS 4 SH 4 7 = 133 ie ° m9 (2,0) = aw POx,y) = 184142 4004+ 402 4043419 + 90243 41409 + 11+ 340429 H+ ore 4+S4 +96 = 18090 ml, = st = o"y? £0ay) x J HL C12 IE KD EK E787 KDA X OD] op LG? xt x29 4 (Par Kort (1x 32x DH 2x 4x 1) J +B x 2D 4 Ca? x27 BDH C33 +024 x00] SCP ID EPR? KBD CH BKODE PK 472] = OSs CR SHB HO + (BLO +108 4 64) 4 Cigti0s + arto) 4 Cbtgr2go4sia) _ 254)80+ 2071+ 720 _ = 1132 Te Overcome Central Memenks& — if we change chape © size cf image, bls Moment vilue cilia change. Ths is the Aeadven tage of basic mament+ This we we, central moment 2 he central moment Fs given by, AUG, py = et st Croat etyag th #Gy) .7 wu(0,0) = a = Flayd = mlo,0) JiCi,o> = ttle 1 =e, ut(2,0) = ML, od~ ximMl1,0) ALO, 2)) = 900.0) Se aon). wd = mas) -y!mu,o0) sp Pan sus aa | “Explain Polygon _Approximection cuits “fey -— > Here we Pac external charecterstes of 97 Imege by sey en-ting He mee — 2 The main technique used here is splitting Eechnique- —2_ Steps For splitting + ") Consider a Closed! aay WeSelect mejor axis he. He farthest point of fmege. 18) Sample the mejor axis iv) Gelect Minor Axis vy) Select points & fasrr polygon - Page No. (J Pathak ts —unitbemy Metory $d _ SE TPs eniteny maton, theo — —*! 7! T wT = TL | Met plying both _cleles by T ——______ | where inctlertes complex Conjugate | Yle lenow thet 2b ~ BPI meets x con be piven cus is L T= ! 1 ' yj cl “J ol ~| to- —|—+# jo- BNE caleu lectiny of “Pre _1Sstemalernel “Ey” for | Janay hit fat T Je check whe then _T ie_cuai'tary mab, we Lneed to check Following rele tzanship— sata : frou 4 [1 oy 4 oars fr joa vif soo 2 toe] toast | fro \ Uboj oa sy 4 | |4 o 0 0 _ = 7 =]? 4+ © oj}, ar — _ {fo + © ee | oO o © +J - 2 — | -DET I an By. of lunrtery refer, ______ pare[ TT _|| foc % (nm) = 21,2, 141,3,2, 1,23 Write 8X8 Hadamard Teanchosm mobic & jes signat Paia_geaph. | fore fas: Hadamard chansPosenr Compute Ha dle mand efeetiny Foren | Hacla mard moactr'<_of orcler 2? can le vecuirsiively genere teal. 102) = Hie) x Wie) His) = Hed x Hee) — fi |__we Enow ste | The _& point _fask Hadamard transform Plocograph— _. $06 shown inthe figure. PageNo| | ca S| (Ss Oy 77 Date | Given Od = 21, 2, 1,1, 3,2,1,29 —__]] whaye TE = Xoo +x) = 142 atoS— ON) = to F vted4+Xlsy = 14 = = 2l2>~-xl(3) = 1-1 =o F alad tA = Broa _ : = UH) xls) = 3-2 =) - - \ = ACGDHXO = 14a 23 = *6)—XOWD = I~ =, a _ e* a oe a Es a ae = _ Me A N= O= P= Stage Me X= T4+M = sa8 =12 - . XD = D4nN = -) 40 27-1 / ; x= KtoO= 142 = 3, — xcd= btP = 142 =4 xX@= 7 s-B8 =-3 Discrete Cosine —Teansform Cher) s - —"bcT_ is uses Mainly in Image Compression we . Poraraaily Uued in _ TPEG fenage compression. _— Z | > Matrie fir DET ic_given by, = oe wo os US ors ee T= ]o0- 853 or270$ ~—- 9-205 — 0-683, Os Ort ~ Or Os 0.2705 0653 9 6S2 0-2 708 01270" - 0270S os 0s OT os o2T0s 9.653 653 W205] [sa 8 6 13 ve GS | fore 24s3 0S PRAT po 3B} 4.347 -V2TOS) J ow” Or2mF “OS rpebys “leo 71 ~1 1S [foe pam pf obs3f isi 0,224 076s 0653 Jf or-S p40 OS gone [12 0007 3 orbs = |-269 -oa4 2.3 o3g | ~ | | 35 146 ike 63 ~ = 003 160 095 yas] Date A ADeT expresses a Finike sequente of date points — —tn_teems of 4 sum oF car ine Fuacttas Oscilertirg ——ork_al'fherent freq uenci's + Explain 2 for aigitel Page heaving alt & pepper Mole, Median Filter fe the best SI ker —_ Averaging filter remove rote by brurciing sb th te n0_ Longer MECN. Put I'n the process, tk also blue _ CY oe Nigger the eweragiing aes ky Pore _'e blurring e || 2 There_ore tres tuhen sage °2ntainy all & pepper p noice. ZF we tide Qveraging Biter Poe .genrou'ng He seme, %E will blur Ahe notre buct~ 2+ evotld ecteg wuUin Abe eg || — Hence _we reed to elim) pate salt & pepper notre 5 _ | we work _eus'te Non-Linear filter Known as Median _ eee er, They. ore Qhio Oreler ctetjctic Filter becuse _ Sher veoponse/s_baced_00_orden'ag er _rargkiy of __ Trels conterinedl .o0ttin neugbke 5) Tn thre cone, we tte a mask _c/milor Te_averagiag 7 filter except that Mask heg no_vatyes do itt ike — | Working oli rectty ob _y_nelghbor of centre pixels J Steps For Medien Fi Ltediags 1) Assume 3% 2 empty mak . _ ii)_place mask _cct_lePt hare _comen ss — 18) Arrange Spinel in ascenctiag. | descending oreter — | jv) Choose meclian from SD vattes. — a Place -pnis median at -the_@ntre = FA) Move mask jo clml tan ~Rechion sto aseaaleg Loner IF we_onty se_move _ no tce_witrout_obistrurtsing edy Ree ec cokes ne S. let_in Fig. below 2-60 _be satt & pepper ole | SLE we use _tverag ny Filter nolge would s/redel out Le eclgew would also end_up getting bluared—_______ |} do jo tote ito yo aoe to to) \o lo yo 10 yo lio 29 10} 10 ‘0 \o lie 10 jo | to oc 1o yo ° \ os = fo fo So BO Vw 1 yo so £0 er 6 oO) 0. oo) 0 so so SD go SS co 0) SO SO SD DD re Wle_ctert_placiag mask At let corner, lAle Moore bo ders. Accarying Q pisets_in Ascending _Oreler weet Lio, 10, Lo, Lo, 19, (a, 19, 10, 270) weeltan = 10 Ais will be pkiced cut cenlite- _|lale new move mak to new Location &. continue Same perocedere,_perForent A, Phis_Operectiin on Image, we Con c30c tat tre rite tt elloairatey without ais bering edges. = PMeclian Dike Bho ‘ostonishing resul® y ever when anmunt of nojse setabively by ges : Date| | Suutihy /Conteactict & Laplacian ix better than gradient etecton oF eclges : = | Feutse. The Line prolate of Laplaeien Opercitor js_shown below: || Since Lopiacian ic cr second dertuatioe Filten, Ib te very | sensitive to noise. Hence ia_a Imeye, DP tiene can -_P0hte. vw 7 entice /naje. Bromdliy deconctly rrqgartade oF Laphret'ar produces __clouble eel yea whith a0 cnctestyeble ePhecks Re crnue_ok-_-fhee Awe draw| fac Meplactnn aa tied fire edge cle tection _ | prevent, Laplactan Gia very Cirge values © ouins Ve s ce ——————— Tenage “Enhantement. technique i9 Spatial bomain _|| Spatiat_ctomcin _meeny tuortiny ja given pace, 7) sthic_coue Aine jangyes Tt Implics worklny volt phyel Nalues or {1 other worl woeklay dlirectty with rea data, | Apatiat_domea'n enhancement On be curried out in two _evceys2 Point Procesing , Neigh Lourbood Roca ay || Point Processing HS we wore vite siggie pixels te. T re IXI operator 2 Nevo vetue of Fay) depends 09 operetor T & present FLY). “Ss Digtat Negative, Contrast Stretehiag, Threshol diy, Grey Level suiting, Bit plane sire Neighbourhood Pro cessiay 3 we Not Only comider a pixel but aly 6b immediate neigh bours+ Wwe C7 eharye veatue of pixel 40,Y) based a Veeluey of (6 &. rely bdsors su shown Irsterd oF 2x2 netghbor te coutd alfo use SKS 6r IAT a _ y+) a [Fema | seen yd | £eeny4) a Seyy-) | FO) FH y+) ae | Fcxayyd] FOanyy | fot gay 2x2 neiyhbourtwod ena Mask. \ || > Te achieve ~Retyb bornood_, Procesiing We place megk On —— the imaye, mutttply eacis component of mask vith | ——COrrea ponding vette oF innage ,cidel therm up & ____ }__place He vetie Har cre pine contre» = -|| ST's opercition it came ay convolution, SEF te -originatimaye. & 9 Li modi bed Image Hen gly = FCA Yy Pee EA SU ae PA ye Dab Wa Pla, yor : WEP Le, yd + We FOG yey tb Wy Flt4, Y= E we f(y) Awy Karty yr _|]_Cenesteete PET ting DLT — Fina Frou. pple oA = C09 EO) p= ftoo~Ffl2)y : te FO Fey a a meftn-Fisy als) = Ate = Ploy+hoy t+ FUd4 FO tO) = B-Ne Flo FO) FU 4 FO) 5 aw = A-c | fod+foy —FOY~ Fa) ALIS Bedye ftH- FO) FO} — FOd$ « We First cateleete Row wie _ a For youl Ao) = OF V4ea4i St a amy ot) = O- @-jayare a ot AW)= oF2-l-l=o —2 A= 0-2 +5-) Aratlarty Calewlertiny Atl Way We ger - e -2 oO -2 Ss 2 mh - ° 12 2 Oo 7} ~XN- Bo -2 o° te - Now cetteetlate Column swhe oe _ For Column 43 me) = & ABHN274P = Br nit = BS +8 = 4 yee _ a mie Fy -8f mien Lalerloting att cowmra we get Date | _ ye re Dee [Oo r-x.94 ys yo) — | ~Detioe + Cuctececn Distonte, City Block Distance, Chess Roarel Disternte, }—Euctedecin Distance as = J—=3-Sbraight line dlistonee bet” two pies |} iF p & 4 ere Je00_pixels with C2, 541 & Oh, hen _ Ms connethyity be (99 =1 eee Blocte “Digtence (be) i 5 for point Pp CH,y 9, lays) Cy block diutace 35 ole Bineet_ aa _ Deglpgd-=—beuaseel tly og al |__ S it meestiree path bet” piel hosel on 4 connected _ nelghporhood» Z = 2 piret whore eclges touch ow Lunt apart G pire | glow touch! ny Aogonatey are ugk spit. a - ee \ 2 | i ) ol | ee _ i os Chess “Bo are? Sis perte8 _ cleFinecl a9 _ Deness (p,q = De (pq) = Moy IXi~xal, { > Measure path bet? pixels based On P-connected Nehohborhood» oe —S Pikes whose edges sue _-forch cre Lunt apare pf pe Pepe fel es J Of fe EP eal _ a Ze 2 2 a 1 -Co Ohectivity & 7 Sect +0_oliminated louble part That Cun'se shee we tue 8-000 nectivity pO 2 2.7) Mere cond” vee ts pixel wb). o 2.0 os belong Ara.region IF batye — J ig 2, When we tue O 272 ole fe} oO & connectiv'ty ue gee ° Lu Piket_cut_ aight Upper Corner »This_amb} gquity ts remned _vsley_ao-connecth We ns vol ce [™1-connectivety ole Mined ay tp XKq_are cai to be | rr-connectedt JF iv Na lp) OR ges ic Ns lp) We set N4 Gp) Nag) ts —eraplys 2 a get_connetted pote[ TT] —||-—Jenaye _“Reatoyertion nage Enhancement __ | Plmprrving tergge to meter =. Hagproerng Lanage To _____||___ fre Onbginal_orey| cline come hidden cletar'ts =| oe Acinzug into evide, O04 J oa | _part_of Lange + Ab ts = — —_ daly a pplicel Locariy fs pot — Deceacanly conghained to__ pk —_ Matcha, original. Peat Jona ge. _| Butt jgbet Looks clearer | 2tompansate for ctata_ > Tea to fer smage to get | —? makes pictire Look better 7 back cto reat, a image | __tulthnoute regerrely st 40 0 || —deeunt Look good « Ze_opyar lé realy uly chould be etarsunin doen wipe | fades (pe-stteted telly but a actual «tertine. veliies befor. ib got clegratled. £ lage 3 fProuate_en'ten'on. of goa 2 manipielates canege that (frets wptinal estinte. to take aadvarntaye of _ OF clestreel resell a peycho, poy Lal_aypects & - human wl euat syste a | 2 “Renpucel of innag nirsik ctretebiy —beyed | applying FBmetivadie DE | peimon'ty on the pleasiay bietering Fanctea iy | eupects ja the viewer ___trerhrection technr'g ws ? alten yitcwal_taypact Hat _enron., notre € geometsia_|___image haa 0? _jaterpreter/>_ _ ais tactitn | 4 Ruthiva drat iimproves tin formect lo _tonten ti Wath chenuferms ty nothing but stequerte ortlered Hladarnard Thans ferns —_n2270tn'Xe oF in_ecch you ZE ux arcane these po F690 change 14 Vis 2gat af motive @re a0t- oh ign changes dam ard free. ULB. opus 2 la treasiay onder get wath _preanphem mats Bin Change | pA) Ge je fe] ela [ | Date, Justi hy | Conbercticrs AN Image compresion technique Ore Paver He Feue dt LE tfp xo) hanshumed by 9 tuachiaa T predwe an ofp Yo then_puttiag yo /0T._/hverse furnctien , ovet peck. produces, ttn) Kb vice verss: Dees Yoou= TLxem] & = - wtnj)= TT Lycox] oe - educh a hans Formcetion Which eg Snvere called || favertote, Nee att functiine have an ‘averme: —— Zr ineye compression , a Freer performing Lossless Compression ewe can yet hack the an\ginat_/maye+ Hence hoscle compreatoa kechnig wey are_toverttbles —rowever while Pertoaniigg. Lossy eomypresston, be _ _Adlctere! nat cate echicn ty not Gkcse esaaley igo bea Senet 1's! mposetble 4p yet hack oniyloet_/nagpe + _ Hence aut conpresrioa technigue are notin verhble. What—are_cli'fFerent oly pes al Reelin clanties® Cocting Redun dancy 2 — a 2 _tode_is_ar_system oF symbolised to present 4 body oF information on. set of- event _ 9 Sach piece _of in fo_or event is cusigned + 9 vente of Code symbols colted 4 eocle word: No of esyarbols (9 each tOde word isi length. J B-bit_codea AShat- are wed 40 represent Losens tes a_most__2% sotenily etreys contain More bits teen That Cire Nnecctet 7a repo revert — Faken tty —> Can be avolded using Hatimen lectins —pre es Enouwledge ca so ag! 2 LFF : a blarrect (0 2 a baleps 21 120 - emply culso duplicate Pass ee ing Bille _| peg fing mej has ___ Trreterant/ Psycho istial Redtunetancy ® > -Hurxn perceptly of Jake taal Loaye norrntitty a ——dloes_not_Pavolve— Buen pixel _vatue in Images J!) _generat_an observer cearctyes Por Pectoral regions mentally conbing tne worth into —recegnleable Orupiigg — = | <> _ Berwin thea Conrelates these groupliags cvi'ty __. “3 pot at 1/; Terap wnat [Gots lies Fpecturctapngy => Becawe pise_of- _commelictec! Spebiatty in formations a | __Pe plicated ta Whe perecrtetio o& coretcred plkebar +3 In vldeo_tequere, temyocanity Csreteted phxes | fo_recluce Eaciesian _| offre quennltatiee cinatys tt oF tn _orderto Complete Image at 2 Jakeutty arrays ey were. ~ | Medan Bie! — wed fo re ctice cate = _Pepher hots + edges are not Airturbed on lured (n ofp -—_a.verey/nj_mask has No . — _|—walites, 12. _€rnphy, ing KK ju _eesed 7 fs _tyeols con _~n_nature__| =) Nonb/near_'a_gatire —_____— Date —Explain_Ia\avelete _Tian.sFoma : _ _==3_wavelete analysts ts_extgtnally hnbodeiced. tn order to |__Improve_seizmic_ signal analysis _-Fe-pee—bel_by Seoltehiny Bam chore Hime fourer analysis 72_____. New better efgedthou te detect << onaly re —arbaupt changes 19 ~fynats « > Fowler anatys's ts based on 77 Muerbiaivety hoag Cosine wate OF 4 specrArc fregventy —S Wavelete analysis te heed 9 chore durattn luquelete of a Speerfie center Bequercy wCs,C) = J Lt) Phe (ED ate Flt) = ff hés,0) Ya (0) oltas h(sc) (£27) ~ ep @ convert signal fora tenes of ewavelea provide a way -for analycting wave forms, bounded Wn bot Frequency & durotiin Alto Signals 40 be stored More ePhrerentty Laan tourer haniforms by better cpproximake. Teal-wortel vignal ¢ be awe to qwelt salted For cipproxiecting ole ewity harp clisonti nut fy Y PageNo.| | Enplain Fiecli lity Colbert + sents — Frectitity orltecta=meacured Io terms of eloseness || 2) Objective Frea'li'bey Cntter'g | bee |0f an _irnige Hoan oaginat source ©F ideal jonaye___ “The_generect _clasies of cx'tedas ee = | 11) —denpective fleditity Co'ter4 | Objective Fed! ii'ty Cr'ter'as —— — when info Loss con be expressed ca —— matnemaetal fof ip € ofp of comyppressfos ——____ precen tbs cara po be base dO Objective Fl. ____ his rook een syuare boror_beb? hws Ionaye Lee fOA, ioe be_itp Vmage £.gley a ep. Jonge | Pesalter QFter com pressiod 2 7 | for cay value ak 4 & y error eCx,g) bet” Ay a flay) glen ca _ _ 002, 4) = Yt Y= Puy) -do__-fotat_error bet” fwo imayes by 2 ee scene Is Lmage tires an ts ain “The _2eok_e0n guar loon) Beta or seen mages ts then ie Sg Wa bed | esror civerciged Ove fVixN cimeuys je __ | [ye eH pe, a aoe re ap ge oe Fee Date| ...] ee ee IF GF OLyts Considered tp he sim bk on'g/aal s'mope = | Foerg) € an _etor or nollie signal € 64,4) te ___ Mean square signal to note _tttla oF output /mage — —— || -elenotect SN Pry cen wgeen MeL, Ned Be i et he _ SN Bny = SRR Se Subjective Aectitity ye ar oe —_dHlere jhe change 1a the sage. _ Aepencts pon She a _hemean _percepher - 2 There are wo —teetanigues_7o_evaluate, She Hange || ta'4ain_dne storage € — eee i) Rating _ tt) Sidle-by = Sile_com, parlor — a erive_—€q” fac_contrask_=sbreteri ey = — Pago No.| | NSH Xr soser 2) Se Sta fP(F-¥) Se 4he-9) 3) S-g = Who-ie) Jed = Pt 2133012), 7 Ss osy OS eg. | abr) } versie Bpomsitei) 4. 14 or sis |} —“By—-pplyres- ave mentioned 72-90 for eettn |e gee re 07 12 03 oy (ae 10 Is 0G o7 12 OF ty Ie I i) 63 os a 3 \ ae 2 foe Puen |e WO 2 BS 3 2 a3 1s | v ge I ar is I ae 7 yo 1S WS aes S13 Se] ’ ee eee eee Wyo 2 = Sind 3 blt TES Codded_ - ——— Calectate Bits per plree (BPD 7 ane Calcdtae Jase &° PSNR oo Mar. yetue 0s S7 Cray Level Bioorg Meelere @ 13 Oot Io] + glo lloe& 12 Oo} [0p é— 13 ooljole Olo saat —00)0Ne—poritt —nay _ S701 Ibppok tr O01N eg Colloo __ 00} lo 9 01010, oot i le 001100, —“brooio oo | Wwe now ofind— -dlecoded j'nrege by Conuerdiey 7 OS |_asrele fo Cleei'’mat Falue._ | prey level P63 Des mal ty a - 13 00} tne D - Sh AU ov SOE | = eemmenli ce Ofc & g [Sb] g a 13 010 wwe 6. fle fel st ie Noo e fs [16] nes Wo. me fe ul ool zg | lo ot | oR ' a Ve plo! 16 PageNo.| | Oi nah (7s 4) + (12-16) ] : =k x Os APIS FOEIHTED ES HUG] + fts-ay® Sure wes Cog 2-13-16 yttco-3 = 314062 - Lassen —Compreniva_ D> Reversible — Loy. Lompreaton a Brreverrsble _ > ree peewee {tency _Contalas | —3 Fecons reseed beat cova NO Hows of nfo cu erace _ Lost of Info hich $a tan ___ Fealpueted of on'ginat | __produces dlishort’en Io hme. bong tapas bl ___| - Mean} cq vere Ener (Mse)=0 | Maes _ret zero — 5 Lou compre seit seta __| 9 Very fangs —)_sulpteble font nt Fi > fie_phoreyraph & satet{te Conypresion : irmaye _comprautos _ => rue man, Arita met SS JPEG, F645 ee || tere Gre fue says to _repracot dats “9 || —_@_saxnsi'ty value “Run Lengin Encoding 2 ane a Used tn bmp File image where dass. _sxptnscnied) — 29 }n Broo oF ao, 3) kncocte eee eae = her. ve clivide coguene oF caty ja fee : @® Leng te of Intensity valu — NM) Absolute _ Set —E =) Meprevent lat ee ing Follows ay _veetye 2 ____ 0 =) end of fre 1 3 end oF Page + _____ move __fo_ pert positon __ = 3-266 9 tet pirel fees adel Boe 44 44% sosst — ut 1 | | _ = = “ co a == 2 4 4 3 3 3 aaa ae Rao Lengo CALYIP24 4333 sp Run Leng tn ending is very —tioyple to use bare __ if cthere yelies jo _olecta_stteean clo nak repeotPoes— her a chame oF yetbiy rin Lenyte onde Qrerkn Athan _ealginedl dake apbon po ! oe 4 _ _ ee ee ee = rt ot a] _ a= RLEMN314| _ Page No,| 1 Auffman. Cocliny & —)_Losslecs culgosit ory — we Find vew table Aengtn code wit min redlundavney Steps TY Qrepare Usk oF probebiuttes 11) Arerange proled bj tiptes in_evescencling Order en it) Adlel Last -heo pre babi Mas - iv) Repeat Born Ci) enti tingle vetuc Founel Vv) Drew a Heffran tree brovergy'n ¢ ae Cae’ | high Prpobitinns _ _ 4) Aastgo Ie Aekt_chitel_& Odo afgor chile Mi) Obten huffran code aaa a a auaccuenees _ —ftete dtr 29 222293993 age eS OO oo ot ne pict hn _ a ai} TF pet ons. deas a 7 a G Ody Oru | Sos one Ue 4 ont v3 on ee |e on | 1! Orbe Date | { a OnsT Ora - LAs. Vis Oo — = Oree 0,2) ae a EES feu] =z - a _ “NS =| for} [ores] = di Pinar Code tutor te ae } QO] ko | —*% ato 2 oO a ee 3 Bu "4 on = joo ie = eee Lo fb on _ te Ato OClo\ a oe ee aa / oO loo _ a : — _ - — = | Arithmetic _ Pa 3 | 2 we Lind renge of, proba bi titra p—» blow Lelue + fog each Letter Or ty. 09 bo} wu) ty —help 6F_foeeban gluco sequence 2 Range oF prbabititias Ns caleute ted oy —tntervat (2) = Lh, b+ lU-2) Pi], Lh tlu-0h,, 2 tlu- BY, - 13 Lower Umit UV Upper bdmip Py Po Pa, Po 2 Rone A pm Lach rte PageNo| | Plo Range Oe . — [0, 0.29 sn fee ce 0.025 0 EEE ojos Ae) \ . — 0.088 «| : $/e|——— = _ \ F 7 : |__.-0.015- — pf —$+-— 0b 62— a pis 91 038 AL: 9.086 | Semper Gray Seate C7ES)e | — Lossy compression techaly ue —_ er eeee ae 6 > And binary vatlee 6& glven Joteoslyy ma give lems A) Lowe bit of mocli Fest plyel aw addea to prevent pret iD New MSR OF presertt pixel Ore Hoken ay TGS WD Yepeat CI & Cit) after moving peeto New _plret | Vv) JE MSR oy ale $f Ve the adel G06 O Saslead of CsB bib: OF previcus ur _— oe aey Lovet Binary Sum _ ze loo OlLo o100 ONO OPO _ 110 Olmo tlio Ons Volto fat O11 Iioo Otnyyto ory - fay Ot |log _fovol OF Os too 0 13° goo voto pooot tt oO 00 | a 1,0 O1lo lio Ol 200 too 1000 si ool ore oc i Be 11010010, tlottfoo, Not Strongly suggest to see Sandeep maheshwarti — Hideos

You might also like