Image Seymentation subdivites an image int its regions on objeds- Beas . Seren tation accunacy — Aekamines the ever] success 07 hile Computerized analysit poncedunes- use f infrared imaging by mniliteny to dete objects — with shay heat gignatunes - Syaea tho conned approach '- CO) Rena Approach ( tarmp. pels dered 09 commen property ) Ww) Boundary Poproach C cata regions that dften in prepents Wke —inkeasity, color eh ) of Vistontinvdian > T has two Dekction (i) Point C isolated point) yw) bine (il) Edge O An isolated point 2 point \ se signih cont nose ney Jeve) & signir ony dont fr is back ground in, 2 Homegenteus aya j= T-| Bale y Rae er fe -Z wa) tomme ¥ Mosk, ta [ele if a Toye f Jal %T a point is dekeced UH) Point _Deketions= how T& a nen a ee Smee, Sample Mask, for post Detection amosneaseaeeseoneeeoeeeeaeeaheaeraitdaLine Delechdy_!— fis type f wnasks ane used to jet responses that & venh'ad, \enizonts!, +45° | gee ag RL RR Ry. 2 AZN E Eee Honizenta | EEE ee AL contain print in image IN) 2a baat yi tat PIN AW to de mone likely asedehed with line iin dinectioy “ mak j. Edae. Deveetion :— pry Gai 4 sek f connected Pixels that Yes on tbe boundarr dy two Meyions whith afer to V7 Valoe. ‘Ly Ed provide an owing ofthe object. n€iyp = Ao edge canbe extnacted by EER + Maninde of derivate + Dinectioo of the denivative Vectn Step Edge. Ramp Edy ¢ Spike Edee Ref Edpe JS yy I lA. fo abaat ne eae Yaak tage oh | rea - dq x te instantaneous chye iv It 9 inknly origin) ies an e Fdensity Tea: distance.etree ae Stages Edge _Dekction:— Stat =| _JEitoig | [Diese Tes teo]__Poirg > eng UL v L inuelves Aistiepshes cua Sorting edge pine bac i: seplicity on fom oho i. Qo imply pivels tage Ede delechisn pigovithen— Denivative Types —a Uses Acfferentation tehniqne fn aye Actetion + Tevphtt Phi + Us femplates | Poet nevmbies et shaped Lash — peyivativs — Very effective fen neal tine Image Patern ft pppsach Surface is peeled 45 hoprgaph ic sunfire with prise] Valve nepneseoting itvde Thneshel ding [TH & on impontemt echnigne fin ienape Sopmentrtion + » TH produces ynirm _negions based on —tiieshold carte’ + from a pyle image, thresholding can be used to cneak binany imye Crlobl Thnesheldty> Lox tray? ‘Tyee \ os ce SL Thnssheldin? co Dpertion depends Upon only weignbrk L "i sale valve: ik a fs ities T depends 0 : pixe! coonytes actourt~- ee ry ee ere er ewe ee KH we ~> Histogram ic used to And — thneshold- Single level Thneshslding : — The objects tan be extnecied Ing peel values wi Heesld acted by compaciing pire! values with Multi level Thaeshelding Le The — obyeck tan be extnached by compaaing pire] Valves with vnvlhple — thvesheld Ty th ---Tm | Edge Linking and — Boondany Delechon. : — —> Set ff pixels Seen edye detecting algorithms , seldom define a boondany ompetely because of noise, breaks io bourndany ete. . There Eye deeeting alyovithms ane ty pralty J, \lewed by Lokiag : 2 ype - Loca) 4 Global Local Processing. — AN points Anat ane simian ane \inked forming a boundosy of piaels thet shane some — common _—pprteperhes- r= Tt is technipe whey P voogh Transhem ‘ an be used to fsolate Features of 4 pantiedlan shape within an inonge Tt & roost commonly used Bn — delection of replay ouves such «as Iines, cindet, ellipses ete + Bys— Tt b taknint of gaps in fattine — beundany descniptions and bs nelytively oneftched 5 imye vaeiye. 7Lingges Y= mRtC into cingle — psint in (p, 8) phne Pe Xo0se + YSin8 + be fa) Oe. a (wed te conned doin edge prints) bea 8) = Hough Transhrm approach i Find the points ¥ Infensection in the = cwnved, each of which connesponds toa line in the Canitsian xy phe. p= x Cop ty sine in cindeé (reat (y-b)* en =>. liane ce mt; _ Steps Hough a Heohane’ AR. 7K » Gee i? Sind Poca maing in Hou Space im paniameler space Region Based — Segmentation : — Rein growing eG see) procedune that groups we inte lange Heqtons based on predefined critents fn Growth - 7 Feng) denote an input ivnage aunty. 7 SOug) denotes a seed cumy containing 18 at lotions - of Seed point. ce 9 denvte —¢ prediste to be applied at each Jochen. fem EF RS one Bame igze