Scribd
Upload
6 views36 pages

DIP Mod2

No description

Uploaded by

muhammedaslamcb05
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
6 views36 pages

DIP Mod2

No description

Uploaded by

muhammedaslamcb05
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
i Module - 2 Maru “pli, ome dewteudnel 7 a PIM fegponied taw Aw Acpasseateck weetowy 7 Ma . ah column ect Vv containing NN elwtnds yo ctencteal ad v(m) = v(vD v (2) v(N) wy A cote vectel 4 Ce J go atte colleoh ow tp y vector~ ; 3 A cow vaceter Covel N clennents duncted Os v(n) > pvc? v(y)-. VO va Pt my ve Lied xn vee oA pw Ww . ‘ Sepreseccdatins ~ nw xtra) = Q-1 3 sz ? F¢ od 3 i) ao. 1 a2 3 3325 ——> ny L 3 ;_ bn Scanned with CamScanner | a motuy Wg Hye MXN fod M Yout ¢ ncotumng ofatnO alive - alu dD Az almin) = ata, e(aa) : a Card atm) a(miy afr) bes Cltumn srcloug - as fe tut) y eC 2D - - xt, carne . ayn), eC) pty BU pee hey, 3 Thi 8% ee qe cao yor fo THR qs 8 Hasiad “y — rel 8 ow" a fe ae seston vitaugell Aig, Axio: © ee eft f the poe atom pay pon ted, AD me {x (vod, alee) read x(m a(mMiy : a Tramaposutew 2 gq (47 2 BT AT (as) : org OT fas” “ Scanned with CamScanner A Aveplity mal TOW & Aas Unsiant cetera alow alagonsl q Reb tiagonal - pacha “fot dhe ne te Pte by te bee usa Soe IO te ty k cus) tr ky te twa) . ~ \ Se _ 7 = - NEN 4 1 ky 4, ho cz [to Cy % Cai) Gy) o & Cua \q er C3 Co) > fate 4 je rv RR Column Uo crcrlar LE 4 panes wow = OK Colunan: a AF esulab’ motun 8 Goepdy Aut the env, 8 not true Scanned with CamScanner Ettalanb — omdtude uo & BY) The peockiet Ra vetudant matin ned net be The prectact ge “oop, matuen Toe pete Ortogeral t_Unitay Mabra —- Aw pets genal Y ea te ov ntas AAT © A “dy A mate wy cnillecl welt oe Lape copal a8 he ynget r Swit a acl . an" ne a Sy ale el aA haf aoe mes oe le eceagene « a wit Blok Marin shoe oe “ block TO Scanned with CamScanner wut Lane > fg a MEK Block, restun’ Auman PE Y: feeptily 9 ag the look Hae Coe eh caltool block. duulont oe fecptity GMs entiebemt = ye (Bloc ey oe e fa f > ceclat * haw toe are Aoadily Hore cheplty Scanned with CamScanner as much as possible PE Gly AMOUNT OF int slumedia and hith-definitin television st decades. Hence. storage mat Won that ty handled by computers has grown exponentially over the is a major problem. The annum aMSSIOH OF the dital way component af multimedia systeans extremely lage Nigh-quatity imac Ht present images at an ace oe of anatiy ' cigth, something which the ey aye space and transmission band- width. something which the curren i and economically. One of the possible SoluGOns to this prob ly and e I « jssion time can be reduced the storage space and trans- For example. if We want to store Mernet, teleconterencinp, e data Hts OF ste idle tec SMS 1 Compress the infarmution so th * 1600s 1200 colour image then the space required to store the image is 1200 «1600 « 8 & 3 = 46,080,000 bits. = 5,760,000 bytes = 5.76 Mbytes The maximum space available in one floppy disk is 1.44 Mb. If we have three floppies then the maximum space is 1.44 x 4 = 5.76 Mbytes. That is. a minimum of four floppies are required to store an RGB image of size 1600 x 1200. This fact is illustrated inFig. 9.1. The amount of data transmitted through the Internet doubles sey year. and a large portion of that data comprises of images. Reducing the bandwidth needs of any given device will result in significant cost reductions and will make the use of the device lam A sure affordable. Image compression offers ways to represent an [ ‘mage in a more Compact way. so that images can be stored in a Fig.9.4. storage space fora colourimage compact manner and can be transmitted faster. LEM muuc3 ‘nage compression is possible because images, in general, are highly coherent, which means that the ‘séundant information. Compression is achieved through redundancy and irrelevaney reduction, Redundaney zis duplication, and irrelevancy means the part of the image information that will not be noticed by the man visual system, meu CFA AUN Reg im sulintey can be broadly classified into (i) statistical redundancy, and (i) p hnovisual rea wel edundaney can be classified into (a) interpixel redundaney, and (b) eoding Sn ancy can be further classified into (1) spatial redundancy, and (2) temporal redundaney. The “dundancy is illustrated in Fig, 9.2. ‘ion, Scanned with CamScanner ‘446 Digital Image Processing ‘Types of redundancy ~ —F Statistical redundancy Psychovisual rodundancy 7? Interpixel redundancy Coding redundancy fd Spatial redundancy Temporal redundancy Fig.9.2 Classification of redundancy 9.4.1 Statistical Redundancy As stated, statistical redundancy can be class redundaney Imerpixel redundancy is due to the correlation between neighbouring pixels in an image. It means tt the neighbouring pixels are not statistically independent, The interpixel correlation is referred as interpn} tedundancy Coding redundancy is associated with the representation of information. The information is representedn ; into two types: (i) interpixel redundancy, and (i coding the form of codes. The Huffinan code and arithmetic codes are some examples of codes, Some codes may b more efficient than others. Codes should be efficient in order to compress the image effectively. 9.4.2 Spatial Redundancy Spatial redundancy represents the statistical correlation between neighbouring pixels in an image. Spatial redundancy implies that there is a relationship between neighbouring pixels in an image. It is not necessary to represent each pixel in an image independently. Instead a pixel can be predicted from its neighbous. Removing spatial redundancy through prediction is the basic principle of differential coding which is widely employed in image and video compression. 9.4.3 Temporal Redundancy : ‘Temporal redundancy is the stati The temporal redundancy ical correlation between pixels from successive frames in a video sequent. © called interframe redundancy. Motion compensated predictive coding is employed to reduce temporal redundancy, Removing a large amount of temporal redundancy leads to ell cient video compression, 9.4.4 Psychovisual Redundancy Psychovii : Y . . er infor HVS, visual information is not perceived equally, Some information may be more important than oth i mation. If less data is used to represent less important visual information, perception will not be aft This implies that visual information is psychovisually redundant, Eliminating the psychovisual redund z «ote a the al redundancy is associated with the characteristics of the human visual system (HVS). It! leads to efficient compression, Scanned with CamScanner 9.6 CLASSIFICATION OF IMAGE-COMPRESSION'SCHEMES Image-compr ied into two types: (i) lossless {iy lossy compression scheme. Lossless compression scheme is preferred in the ease of medical image com> pesion, Lossy compression scheme is preferred in the case of multimedia applications, In the case of a fasless compression scheme, the reconstructed image exactly [Link] original image without any Joss of information. But the compression ratio that can be achieved using a lossless compression scheme ts tuually less. On the other hand, a high compression ratio can be obtained in a lossy compression scheme Sf the expense of the quality of the reconstructed image. There is always a trade off between the quality of the reconstructed image and the compression ratio. jon schemes can be broadly Jompression schemi 9.6.1 Lossless Compression or Reversi ible Compression In lossless compression, the image after compression and decompression is identical to the original image and every bit of information is preserved during the decomposition process. The reconstructed image after compression is an-exact replica of the original one, Although lossless compression methods have the appeal Faerie Mp deterioration in image quality this scheme only achieves a modest compression rate, The Ioatleve compression scheme is used in applications where no loss of image data can be compromised. 916.2 Lossy Compression or Irreversible Compression intosay compression the econstructed image contains degradations with respect tothe original image. Here yore reconsruction of the image is strificed by the elimination of some amount of redundancies in the image to achieve a higher compression ratio In lossy compression, a higher compression ratio can be we ined avhen compared to Tossless compression. The term “visually losses i is offen used to characterise lossy compression schemes that result in no visible degradation under a set of designated viewing conditions, Scanned with CamScanner Peet SSeS) rl ge-c standard is to enh; The objective of an image-compression si > : a among systems manufactured by different vendors. Since the mid 1980's, members from both ITy and TSO have been working together to establish a joint international standard for the compression of grayscale and colour still images. This effort has been known as JPEG, the Joint Photographic Experts Group. Officially, JPEG corresponds to the ISO/IEC international standard 10928-1, digital compression and coding of continuous-tone still images. After evaluating the number of coding schemes, the JPEG members selected the discrete cosine transform (DCT) based method in 1988. From 1998 to 1990, the JPEG group continued its work by simulating, testing and documenting the algorithm. JPEG became the Draft International Standard (DIS) in 1991 and International Standard (IS) in 1992, The JPEG standard was designed for compression and recovery of still photographic images. The standard provides a relatively good compression ratio. The most commonly used transform for image compres- sion is the discrete cosine transform. One of the major reasons for its popularity is its selection as the standard for Joint Photographic Expert Group (JPEG). The JPEG standard was designed in 1992 for compression and reconstruction of still photographic images. The JPEG standard describes a family of image-compression techniques for still images. Different modes such as sequential, progressive and hierarchical modes, and options like lossy and lossless modes of the JPEG standards exist. The JPEG standard employs block-based DCT coding. The meaning of block-based technique is that the transform is not applied to the entire image at a stretch, but it is applied over fixed blocks each of size usually of 8 x 8 or 16 x 16. The basic encoder and decoder structures used in JPEG standard are illustrated in Fig. 9.38. In Fig. 9.38, FDCT stands for Forward Discrete Cosine Transform, and IDCT stands for Inverse Discrete Cosine Transform. From the figure, it is obvious that the input image is partitioned into an 8 x 8 sub-block. The FDCT is computed on each of the 8 x 8 blocks of pixels. The coefficient with zero frequency in both dimensions is called the ‘DC coefficient’ and the remaining 63 coefficients are called the ‘AC coefficients’. The FDCT processing step lays the foundation for achieving data compre ove» oY concentrating most of the signal in the lower spatial frequencies, The 64 DCT coefficients are scalar- quantised using uniform quantisation tables based upon psychovisual experiments. After the DCT coeffi- cients are quantised, the coefficients are ordered according to the zigzag scan as shown in Fig. 9.39. The purpose of zigzag scanning is based upon the observation that most of the high-frequency coefficients are zero after quantisation. 5 ce the compatibility and interoperability Scanned with CamScanner Image Compression 499 » 8 blocks ax8 DCT-based encoder Input | FDCT >| Quanise of |,| Entropy Image 19220 encoder Channel DCT-based decoder |, [Reverse |, | Entropy IDCT 4D DCT ene zigzag |“ | decoder Fig. 9.38 Baseline JPEG encoder and decoder AY ‘The resulting bit stream is run-length coded to generate an intermediate symbol s* ‘ymbols are Hufliman coded for transmission or storage. _ To perform JPEG decoding. the reverse process of coding is applied, that tntropy decoded and passed through a reverse Zigzag table in order to generate @ Tescaled using the quantising factor and then passed through Inverse DcT. Fig.9.39 Zigzag scan sequence, and then these Reasons for Block Processing The reason for taking DCT in block manner i 1. Taking DCT for an entire image requires large memory, hence it is 2. Taking DCT for an entire Image is not a good idea for compression veithin an image: Scanned with CamScanner 4 490 Digital mage Processing e functions, There are four types of AO ee Itt and DCELV, derived according to the varying boundary cov : namely, DCTH, DCFH. DE u viunage and video coding. The 2D DCT is performed by applying 1D ney DCTH form is commonly used ak © DCT is an orthogonal transform, The DCT possesses 8004 Varianeg on rows an columns separ ly y compaction in the transform domain. The roletot DCT is tp siseibution which Fes 1 et nd ae components and the role of inverse DCT isto reconsy a1 into its a tions, The decompose the original si a ¢ signal. . ifs " The 20 DCT of the input imge fim) is given by FU 1) as Be [Qu Dnk) (Qn nt FEN = alka) OY Sm mpeos| S| HW (ox momo where a= (9.23) and . a()= om Advantages of Block Processing The advantages of block processing are summarised below: 1. The reason for block processin, complete imag. 2. The pisel correlation will not exceed usually 16 or 32 Moreover. the correlation between pixels is less bet pixels ig is the transform of small blocks is much easier to compute than the pixels. But this assumption is not always vali. ‘ween distant pixels than between neighbouring & One of the drawbacks of DCT-based transform: is caused by the discontinuities that result Drawback of Black Processing block artifacts. Blocking artif of the image data Minimisation of Blocking Artifact The blocking artifact can be minimised by the use of (i) overlapping blocks, (ii) low-pass filtering of boundary pixels, and (ii) lapped Orthogonal Transform (LOT) The use of overlapping blocks results in increased bit rate, achieved, Low i and hence hi -pass filtering of boundary pixels will r. esult in blurring, Scanned with CamScanner compression Metrics The type of compression metres use fr digital data varies depending on the type of b 89 97 == rie ~ Scanned with CamScanner UU 2p ALE aise slit ime! oft cconlt > Gn AD Ave j covclotion bf pinets nw a sconlud ua wulyg ° ; ces castor gr 2p REC redler b&b ache é hed cert biow blew pois ai mange Crouuag Ay ce also utelyed pow 2 meus : wrng RL ce at fe “ artrendad Y a on : mo: tos a f oh cds vada Hee 2540 tee gab on : tol pte vO ee fe actual op te te Roe AE am ret gb te Up ote ek ae f oe ' to, tr IE aw Y yo 7 Ol ed om the pre geroal arenes . . appoacks ; New, Ath ve a Aveo} ~ an m-bit weno AMON Scanned with CamScanner > & te Waal outed te St es ig te riage) Ret RH te pot a) | tt om Abit a he Aowut acre Ait Tate Oa ol tang te ae 4H o epee BRT am AM - : uo ‘ C Il ois titl) aga pre & ate 1 iag{ 19808 269) 5 agent fe OTe a “ ad (A j4oo) 4 wo dw $ ° a opocce 7 }1900.009° Scanned with CamScanner ye dosed — . duvclomund af etorly oT ao oe a a het aie oon we oun pian conten: 4 Preciiten 7 pens 4 be! Sa (Es eS dict pete Ve > oclet fe et fo Ee loot ete [oficents Scanned with CamScanner tn) el 4 Srput Conrpressed ee &) Loe an qe. Pre dtetoy [Heaxat Loy [ wateges | Aly) (ay faced a Rtn) Docowsprece sol pet Ey tora Aw) 7 (b) Arcodere juve. : wwodel tH A Aesslemr ne a Loss Perel trod & chase A byer 4 added + de a“ ann - Ae amwoe tae tacked Ps acu ‘ compression” paperenn of wap JOSS not 7 : zematpelts Arnets dah the He a Agceetyde aces 3 om ode > fossy, eacodu's yee plat, he Td eu I Pe hep Arp a eet pot petlicine pte oe eed Alo) + é(ny 4 Ale) Scanned with CamScanner Spr iy p—— Compre Rigs | . as potas, (EL 6 UES ont, HM 6 (#0 o se Aten) aad seo 4 cqe XN) fle 2p PFT foro kgeee d end xd Ob . j2mmk -jzen wo N-loN-l , eon e plat) = SS be" meo nee ro Scanned with CamScanner ke aur 20 peg wv ee oY t j 2a. 5 c(h 1) Jae" ed foo TM Ar) 2 1. a> EKO 8 gem by E(k © RIE wr HO at gop RCRD > Aol P bo x40) 2 rena p 3 eT e cue S Magitede Ae pre byte «gr (e.0)) Jeceol = (* > tha apetem ° age fork A)} g(a) = aD fe (RAY Me AO] et a ap Fiamaper fo OS ig - (i) & a 1 F (nai) year a geperalé to (a) Compaletin PPT “9 w tot a Fr g biesniiai acon ‘tele pg 1D Arefoter Scanned with CamScanner = — Wis wha gotrealy dusted atove + ap eT 8 equivaleut Av preformed -+wo 1D dramsform ad: foun a wo Hamyfeinn 9 20che mee a tandige wo fap gy Mem dor? é de arforra ov 2oth, Column 4 Fea, L) @u (OA 7) aoe dav a ft ope fone) . wd = vn he ytd yun 4 ange f ; a sre ign Aree naes 9) oF FS Elmo mere mea ne? a ne: x dems pen") of mo . . a oy Renal PRI ae fe oe ap ort qe ne 4 ° dee patty a Per ts ) . p(at) — F(A Y 0} q . sible ~ % tomatic FEE wi oir ~ Most porte aposel te egy ge 5 Convo doy py done ai 7 coke aw Fe . 3 loweltiom x. z tk) At ) a , aly) ® at . ' baci g ay . & a : . sap BE oar (my Aiea) “¢ ae Scanned with CamScanner min) Alm yas cbt) fnew aOh Ft 4 bem tla, Ub) per (4 (err) ° ian"! 41) Cons “Efe The ver] €'Cn, 9) - 0 ga ey et (4, Flat) =P (AO Scanned with CamScanner P aig fotebon : 2 we ie “ ‘t eae loietd oe et po by He ’ | oa > (sear, Aan 0) - DFT, if ( Aco8e towe) | —> e[r oss , ron 4] ves FOE (016) [Bax 4 +H, Ram (G+ 4-)J ali, trap onptct vel Op octhonemal sapere te trod’ oo 7 ” : seed “le apectentt eae bn Auenrs 7 7 poe a j ity et it, i oo The compe oe warren a im be Oly Ho ~ ki au ee by: a usu “ue ; a Tre) te i{n. mola) wal te mel ey) bil = WD ! quit) ~ jz pe ord fon Nv a wor x = Aer Rao HV peanfoors qe r beth pal e L ce fo) Te bile) bp-a i! N mn ico adomrorc Prowlfrrrs eal #0, penn Inconel tee ot ladon ~ ke he 2 Faomsfor™ ated - pass certo vat ee ot we The eet Poi of He Scanned with CamScanner Scanned with CamScanner Riz, compete : mt py f ‘yl a ne Aegan 7 * Deter pe nt pepe 20 KY=! hte ea aak aw) a pte; et a os Daud Comers rent consti Qe wR OBS aman fray SE . compress . > gag # dh > ve fee Scanned with CamScanner whee wt zak y[EJ> S rte” bho ot j good % pst © vars 4* mga) wh ep eh ook Re ad anat 1) AR | cos Gene) a aE aeons? “(Roy a . ECA Rd= wha «(8 = ie 4g, Kee Je * Keo MW, < CA) = [a qe = y Are A roms ww dis crane rospee oy Me 2d eran viet = fee) =o (Ayre 20 Ateais) * gen _ gor x kee Trantor 5 * Sv> oe ol Scanned with CamScanner 4.13 _KARHUNEN-LOEVE TRANSFORM. (KL TRANSFORM) The KL tran + Kari Karhunen and Michel Loeve who developed it as a series method for continuous random processes. Originally, Harold Hotelling studied the discrete formu, the KL transform and for this reason,the KL transform is also known as the Hotelling transforma ‘The KL transform is a reversible linear transform that exploits the statistical properties of a vector rep. transform are orthogonal cigen vectors of the covariance matn ct A KL transform optimally decorrelates the input data. After a KL transform, most of the “ of the transform coefficients is concentrated within the first few comy Property of a KL transform.) ™ ts nan sentation. The basic functions of the KL. a data set rey ponents. This is the energy compaction re Scanned with CamScanner Drawbacks of KL Transforms he two serious practical drawbacks of KL transform re the followin, i. AKL transform is input-dependent and the basic function has to be calculated for each signal model on which it operates, The KL bases have no specifie mathematical structure that Jeads to fast imple Mentations, fi, The KL wansform requires O(n?) mul multiplications. ly/add operations. The DFT and DCT require O(log’) Applications of KL Transforms () Clustering Analysis The KL. transform is used in clustering nalysis to determine a new coordinate system for sample data where the largest variance of a projection of the data lies on the first axis, the next largest variance on the second axis, and so on, Because these axes are orthogonal, this approach allows for reducing the dimensionality of the data set by eliminating those coordinate axes with smvall variances, This data-reduction technique is commonly referred as Principle Component Analysis (PCA). Scanned with CamScanner 204 Digital image Processing (ii) mage Compression ‘The KL transform is heavily utilised for performance evaluation of compress, algorithms sinee it has been proven to be the optimal transform for the compression of an image sequence the sense that the KL spectrum contains the kargest number of zero-valued coefficients. Scanned with CamScanner Bac SINGULAR VALUE DECOMPOSITION. The Decomposition (SVD) for s nin 1874, and extended to r © malrices was discovered independently by Beltrami in ngular matrices by Eckart and Young in the 1930s. 4 decomposition of the form (4.94) Where 4 is anim x n matrix U. Fare orthonormal mai _ Disa diagonal matrix comprised of singular values of 4] {The singular values 6, > oy... The singular values are obtained by tak Equation (4.94) can be written as 2, 20 appear in descending order along the main diagonal of D. ng the square root of the eigen values of Ad" and AU. | — (4.95) (4.96) (Tie relation between SVD and eigen values are given below: A=upV" Now 44" = upy"(upy™)" Af) ae EEra] Be , > J) 24 af et * ah i . 31s 3-8 ‘ee ie a} Scanned with CamScanner a ped tp Wali bans fer foe, 4th euolic P aon (8° ) a md by) bat -i (W) gomed « 5 T ©!) Scanned with CamScanner ! yo b (1) 2 Dba alt gL cypil be ebb) be D 1)" @) * 1) ' ? 4 4 f oxo We sg for *eP s Leite 7] ge here g 4 ive athe), _ jo wy af forte 3 sage ecrps gir hy re (8) Ly penton gle) 7 5 hee WOE ge eh 2!) j qe ° [xo . ene ® eD s \ | 4h eel 2 ey Y ‘ beter all valay } tale transfor 4 coletated - Scanned with CamScanner Horn, fu bow pe nef vi ye Pe Ye Ye Me git) ye My “y Me Vy Ne 1 N "=A -Ye Scanned with CamScanner Tr anasfotin codhiog Image Compression 487 sa ~aransform coding is the most popular method used iy j tt med to produce a set of linear transform coettiei in prteansmission. Thus, transform coding. is imayge-coding, ‘Mts, Which applications. An ima, ' # mathematical operation tha cone a eon and ged pels into a smaller sot of uncorrelated coetieiens form coding reli risa large set of highly ron anage exhibit certain level of correlation with their neighbouring pila. eset ce tbat els inf mage compression. It is used to transform the image data in te nose (© forming the data into frequency domain, the spatial redundanci ra age of using transformation is that the energy of the ransformel y region. and is represented by a few transform coefficients, ‘Read without significantly affecting the reconstructed image qual senal is segmented into non-overlapping blocks, and each block is expressed as a weighted sum of discrete saris functions. The purpose of transform coding is to decompose the correlated signal samples into a set of spoorelated transform coefficients, such that the energy is concentrated into as few coefficients as possible. ‘The block diagram of transform-based image coding scheme is shown in Fig. 9.37. re is linear ane e ly ration is a very useful to frequency domain, By time domain can be minimised. The data is mainly condensed in the low- hus, most of these coefficients can be ity. Ina transform coder, the discrete data tunsformation The transformation ‘reorganises’ the gray values of the image and thus changes the carelation properties of the image. The transformation compacts the image information into a small :~"~aber sfcoefficients. For efficient compression, the transform should have the following properties: 1. Decorrelation 2, Linearity 3. Orthogonality peorrlation ‘The transform should generate less correlated or uncorrlated transform coefficients 10 achieve high compression ratio. fficients. L rinci to“ between pixel values and transform coe! linearity Linearity principle allows one-to-one mapping nal transforms have the feature of eliminating redundancy in the transformed image. Otthogonality +Orthogot Channel Entropy Original || transformation Quantiser oe came image Source encoder, Transmission channel Inverse ‘Entropy, ou Recor Inverse quantiser decoder structed | transformation image || — gore cote 9.37 Transform-based image-coding scheme Fig. Scanned with CamScanner 488 _Digitallmage Processing Quantisation Quantisation is the process of reducing the number of possible values of a quanitity, thereby reducing the number of bits needed fo represent it, The inge transforms by themselves doy not achieve any compression, Instead. they represent the image in a different domain in whieh the image data is separated inty components of varying importance. Quantisation is basically an irreversible proces Entropy Encoder The purpose of an entropy encoder is to reduce the number of bits required to represent each symbol at the quantiser output. Commoly used entropy-coding techniques are Huffman coding, arithme. tic coding, run-length coding. ete. Entropy coding is basically a lossless coding scheme. The output of the encoder is basically a bit stream which is transmitted through the channel to the decoder. The channel is usually assumed to be lossless if one concentrates on source coding alone. If one is interested in joint source and channel coding then the error occurs when the bitstream passes through the channel, and this should also be taken into account. The decoder basically performs the reverse process of the encoder to get the reconstructed image, Scanned with CamScanner

You might also like