fiil-.ffing process iy perUirmed on the ne i qhbor.i ng pJxols by ij^ing a b.i''iine,^r int^niolatioEi filter in the case whero the v^i I ue calculated by usiiiq the sample of tho ne.i qhboring pixels is lower than (.he set threshold v^alue. 5
KFFEC'l'iJ OF THK INVENTION 100271
ftci^iording to the present, disclosure, it is po.'ssilnle l-o encode and deoode an imcige . Particxi J arly^ i t i5 10 possible Lo suppress a deteriofiction .in image quality.
RRltilF DESCRlfTIQN OF DRAWlNC^S J0023J
Fig. 1 is (i diagram illujiL. ratiny rt con^iJ quraHiOii ISJ e;^(imple oJ a codfrtg unit.
Fig. '?. is a diagram illus[.voting an example ol angular pt edict lost.
Fig. 3 IS a diagram illusI.vsiting an example of
planar prcdictioo.
20 Fjq-4isa diagyaiii illuMj. Jiating ait example oi a
fdost probable niOiSe.

Fiy. IJ 1 ^] id ditiqr.Tiii illustr^T l.ifig an o->^nr[Lple of MI>T.S (f^ode DcpnncteiiL liftra ^ityoothing) .
[■'i fj . ti 13 0 dLagiaiii i 1 1 usLratiifq --m exavtipl r. ol a bounds ry value r^mdol-liincr prrice^si .
I'i g. f Is ii iilc±grsin ill njjl.ratincj ran example nC c± decoded Lmago,
t'''ig . fl .! jj .H diagram i J.iustrat i ny another example of a dccnrJE^il iinaqc.
Fig. 9 i K a diagrtun iilusf ral.i sig an o^^irrLjilt? o± a behavior of PI Lhreshol d value dotL.ht?r
example of -d. flow of a threshold value Jsel.L.ing process.
Fxii. 17 la a block diaqraEn .inii.'-il r^Ling another
contiquratf nn example ot a t lire oho 1 d value hit? (".Ling unit
and a EilLering processing uniL,
10 (■'ig . ly is a flowchart i 1 1 UKLrating still another
example of a f ] ow tyf ?i L.hEt?.'^hold value setting prorej-^s,
Fiq- 19 i i^ a block diagram illustrating .s(.lll
another cnnEiguratlon ezartiple of a l:h(-esli(>Ld value
scttiny unit and a filtering proi:fSHi[ig unit.
.1.5 Fig. 20 is a flowchart, i llu.sttating still another
example ot a flow of a * hreriEuild value setting process.
Fig . 2i is a f 1 owch;jr(. i llu.strating stilt another example of a riuw ol a threshold value setting prnt-.i^y.!--^.
Mg, 22 is a tlowchart il.Uistrati riij .■s I i M ^ncLher 20 exampl e oE a flow of a threshold val i.)e I^P.I;*.! ng prcuzes.s .
Fig. 23 is a block diagram i 1 lu^:LraLing a main i:(jn£iguration exampl e of -JJI i rcL-age decoding device.
Fig. 24 is a b.lock di;jgra[[L il lu;i Lrating a inain configuration exanfpl c of a thue^hold value setting unit 2b and a filtering proresriing unit.
h'-i g. 2h is a flowchart illustratinq an exampl e of a flow ()£ a decoding process.
Fig . 2 6 is a f J owchart i 1 1 ii.'U. ral. i iig an exampli? of a
flow of a threshold val ue sE^Ll.i ng priKiE'hi.'s.
:5n Fig. 27 is a f 1 owrhar L J 1 lustra Ling an example of a
flow of a predi ct i -jn pl^^:e^3^3.

:iP."i.'>1032WO00
L'\ig, 2S is a f lowcii^ rl. illustrati ng ^ii t^iOinple of a flow of ijiri Intra prediction process.
h'Lg. 29 is a FTov/chart illy3tv":-il-i ny another cxsinpl e
■ of -1 flow of a t-hfeshoid value .-ieS.ling process.
5 Fig. 30 is a block diagram ill^jstratinq another
configuration example of a threshold value isel.Ling (lait and a filtering procesi^irig unit.
h'-i g . !11 is a flowchart illustrate nq stJ li aiiother
example of a flow oT ^ threshold val ue setting ptoce-^y,
10 Fly. 32 is a rSotJehart illustratJng stiil another
esaiLiple of a flow of a threshol d v^l ue setting process.
Fig. ^~i Is a block diagram illustrating still
another configuration example of a threshold ualue
setting unit and a filtering processing unit.
1^ Fig . S'l is a fl o^jchart illixstvat i ng still another
f x^rtiple of a f 1 ovj oL a thresho] d value .setting process .
Fig. 3b i^ a block diagram, i 1 1 uj-M.i-atiny still
another corifiguration oxaTiipl e of a tlireshold valun
setting unit and a filtejlng processing unit.
20 Fig. 36 is a fl ovjchart illustrating sti 1 I. another
jiX.implc of a f 1 ovj ol s threshold val ue seating process. Fig. 31 is a diagram ill us»-raLJ tig sn examplG of a iflulti-viewpoi nt image encoding scheroE".
Kig. iH is a diagram i11UHIrating a main 2 b conf 1 giira tion example of a HIM J (.1-viewpoint image en cod i rig dev i i:e to tjhich the present technique is appl ieij,
Fig. 39 is a d i ayram illustrating a mai n
oonfiguration exampl e of a multi-viewpoint irn^ge decoding
device to which l.he present technique is appi i ed.
30 Fig. 40 is a diagram illustrating an exacnple of a
hierarchical imaye encoding scheitie.

12. \3
l-'ig, 41 i s rf diagram iliustratinq an c>:ompl c "f spati al H<:ialrfbie eficoding.
Fig. 12 is a diagram illustrtit i ny PIJI eSrimpIt? ol
L.emporal scalable cncodi nq .
b Fig. 43 is a d.i aqrafi i 1 1 vifjL.rfJl.lng an example o£
signal-to-no"i np rpil.iii example ot a hierarchical image cncodi ng
device to which the present tcchiri'lue is rtppli(?d.
1 0 Fig . 45 is a diagram ilJustral. i ng 3 "IM.I 11
contiguration exampl e of ?! 1L Lerarchical image decoding device to v;l"ii <-ii l.lie pre^ienS. technique is applied.
rig, 4n 1H a block diagrain illustrating a Ttiain
coiifi gu tat ion example ot a computer . ■■
15 Fig. 47 is a block diagram illustrating an cxarnple
(j£ a schematic conf igurat.i on exampl c cf ^ l.el ev 1 H i ciri apparatus.
Fiq. ^d ir- a btock diagram illustrating a schematic
confi giiL-nl ion exaraple o£ a mobile phone.
70 Fig. 'ly is a block diagram illustrating a schematic
configuration example of a record "i ng/reproiisiri ng apparatus.
Fig- '^O IS Pi bWn'k (iiayi.ariL J llii.'sLrrj l.i;iy a si:heiiia Lii:
conf "i guriil-i on eKaciipLe oE an imaging apparatus .-
25 Kig. ^1 is a block diagram illustrating an example
ol Qsft of scalable encoding.
Fig. 52 is a block diagram illustrating another example of use of se'iilablc encoding.
Fig. 53 is a block diagram illustrating still JO another example of use of sea]able encoding,
Fig. 54 is a block diagram illustrating a schematic

SP35i032WO00
<'(Hi£igura"ti on exaniplE^ at a video sf-\--
Fig. 5^ fH a block diagram i I Lu^l-.tating a schomatit; conf igurtiLi on exatuple of a v.i dco prtjcessor.
I''it}s b(f is a block diayrr^jii lllustratlnq aoot->if?r b scIicraaL-i c configuration oxampli^ ot a video proee.'j.'-jor-
Fig, 57 J 5 3 di-jytaiiL illustrat.i ng a i^oriEiyutatlon oE a conti;nt repr()dLH:irsg system.
Fig, !3a IH a diagraiti iilu'StraLJ ng s llovj ot data in
a contcriL. reproducing ays tern,
10 I'i g . Jjy is a descripti on lUrigram. illustrating a
^pec t lie example of an Mi^D.
Ficj. 60 is a riincl.ional block di^grrtm i 1 IUH[.rating
a configuration oE rt content server oT ri content
reproducing system.
lEi Fig. bl is a functional block diagram illust rati ny
a coll Ely ucat ion of s (-(inL.t^iiL. reproducing apparatus cif a conL.enl. reproducing syj-^teiii.
Fig. 6? is ri functional block diagriini Lllurttrating a conf igora Men of a content scrvf?r oE a content 20 reproduci iitj aysl.eifl.
fig. C3 is a scq>.icnce c>ini \ J llur.trating an ezaraple of a ciommunicatiofL process oT edcli apparatus in a wi rel (=:.^.^ coininunication sysl cm.
Fig. G'l is a sequf^nce chart illustrating an esampl e 25 cE a communication prrK-ess of each apparatus in a wireless communi i-al.ion system.
Fig. G5 is a schematic diagram illustra* ing a conf igurat "i on exaifiple of a frame format
transmit ted/rei^eiued in a communication procf^s.s by each JO apparatus i n a wireless coicmtjni cation sysl .em.
l''i(j. (>(, is a sequence chart it.l us UraLiny ^n example

(jf a coiri?pUiriicati"Ci procews of erfi-Ji app,di-atu3 in 3. wi rcless iiortcviunic^t"i.on sy,'3tcm.
MODE, KOK CARI"JVIWC OUl' 'I'llE If^Vl':^JTIO^J 5 [0029]
Here "i natter J Gmt^odimciits for iiiipl cntenting the
pj-cisent disclosure (hereinafter, J-fiterrt^d to odimenL"-3) wil' be deisccibed. iij addition, the
dej-tcription i.s perrofrined \ n the io-Llovjing order -
LO 0. Overview
2. .'iocond Kmbodiinciit [Imaqe ijenodifig Device]
[il. Third Enibodintetit (Kuli.i-viewpoint I;n,g^iG
Encoding/Miil ti-Vieivpoint Irriage Decc.diny |]ovii:e)
15 4. Fourth EmbodiTtient [HieraiQhicai image Kncodirig
llieraichi cal liniiqo Decoding Device) 5. Filth [■ImbodiHien.t {Computer) b . ^'x^mpi^s of Appl i cation
7. ExarnplG or AppIJi'stion of Suitable Eri,coding.
70 3. SiTith Gmbod-imenL (.'jOt/Uni t/Modulo/Procossur)
9. .■^eveELth Emboiliment (iCxayiipl o t^r Applied*, i.on of
Content Kt'produ<;i nq Sy.'stom of MPEG-DASH]
10, [!:igh<:h J^mbod i mont (KxartipXe ot ApplicaH.on of
Wirfiie.sH Commuiii cation 3y& t.em ot W i -Fi :^ tandard)
2b r0030J

UoieiurtftGt, ^ho pre.r^cnt i.^cV,niquc vdil be described in ^n exaiHpl o where the VireHent tec:tiniqiie i^ ijG applied to imayc encodingA3(^(todinq in dii riEVC ([ligli !;f ficiericy Video Codintr) st:hemG .

[0031]

^■n the r^VC ^P.d'Jrincd Vi^ltjo C^diEi'L^) sc^e-ca*^, a hierarchf cai sS.rsjcture htiving ni'JCTroblock.'s and 5 Eubiiiai:i-t)bl ocks arc defined. HuwevGr, l-tiF: macrobl ock of lis >C 16 pi^eiH 1 ^ noL nptimai to a large imsye fraiut* called LMJD (Uil.rs Iligli iJcf ini I.J on, -IQCIO x 20UO pixels), which is to be a.ri object oE o. iiexl.--gcneral.i cin erfcciding scheiiif!. 10 [G(l.l2"f
On the otht^r hand, in the HKVC &chf;nio, a.s illusttriLeti in Fig. \, a ^:odi\\g 'iiut. ^CUi i.s de.Eincd. [0033]
The Ctl -i 5 ^isn called a coding tree block {Cm) and 15 if. a pairtiai rc^gion uf jjn iiiiagf? in units o£ a picture
which lijis the .'sarnt? tunct.i ons a>i thiat ijf the fiiacrobiock in
<.hf! AVC scheme. Thet latter is fixt^d in sii'.e oi id x 16 pi xols, biK. the former is not f ixei^ in si^-e but J t is df^signatcd in imac|e nompresK i on information Sn each 20 .sequence.
f'Or example, in rj sequt^iice parameter Het (SPJ^) incliicJcd in oiil.put encndod dalo, a m^^imum ^]i:^G (LC'K (Large.-.t Coding Uiilt}) ot tht^ CO and ^ minimurn size [3CU
75 [SiiiaileMi: coding Unit]) of the CU are defined-[003^i
In edf:h LCEJ, l.he CO i:an be .■split inlo CUs having ^ .■sraaller si/e by sel-t.iiig spl i t-tlag r- 1 wi thin a range where the size of ClJ is nol sitialier than l:ho size o± ^CU.
30 In j:he example of Fig. 1, l.hc sise of LCU is 12y, aud Lh/? maximum hierarchical depth is 5. When the value of

:^plit_^iag '^ s "1", ^.hc CU hrtv^ng a ^.-iW.C oS: 7.ii x 2S& i?
;ipl it into the Cljj, having ^ size of N :i N which art^
dispfsed J n the oiu? lower 1 ayor.
[0036J
5 In addition^ Lhc CU i^i .■split Jnto prE^rJi ction units
(PUri) a 13 the i:"<:?gioji£ {partial i-pqions cf an im^go in uniti-i of a picttif-cl whii:h bocoiiie a proceni^ing unit ot intra ot -^nteE pr^dictiiiMj or t^it^. CU it, E'.plit iuV-o L.ransforin units (TLJ.-i) as the rogiony (parl.ial regions of 10 ^n image in units of a picLurc) which bectHiir: a pr(ji::r:'?5iny uniL. nf octhogonal tr^irisform. At pre.scnt, in tho IfEVC KchomcJ in addition i.o 4 x 'J orttiogi>nal trrfni^form dnd G x 8 in Lhoqonai Lraiisiortii^ 16 x 16 orthogonal [.f tinstotHi and 32 X 3? orthogonal tr^n.-iforiit o^n be u.'sed.
ijiko tlfie above-describerS J[EVC .■^c^horne, ] t is coih.'ii dcred th^t, iji (.he encoHi ng si:hemo where the i'Xi is detided and vrfrc ous procc^se.s arc performed i n unH.K of the CU, a macro!)! nek in L.ho AVC ^ichertie corresponds to the 20 LCU aiid a blook <=iubbloi-k) correp?pond;i to the Cdsreij LhaL rf rnotioii compeii.s^tion l>tork ITS "till: ftVC st-heiinc ccrtrij^poiu^:-^ to t^ie V\}. Hiiwc-v&t, sinr:o the CU has a hiorarchJcal stricture, a size {for
example, 1?Q y. I^R pixelhij of the LCU of t!ie uppcrmoKl: ?5 layer i 5 cfen^iril ly se[. to be 1 arger ( havi thf size of the macroblock ot tht^ AVC Kchcme. [0030]
Therefore, heroinafI cr, the LCU i.'> assume^d to JTiclvide tilt: rtiacrtjlTj^ oci^ in V'-ht RVC scheme, ijnd \M^ C\3 ^s 30 ti^.'^vtiicc] to include J.hc J?Jtii:lf (suhi>'\ ock) of the TiVC sciii^rnc, Haiiiely, the "l:>lo<^k" iiycd foi the following description

i:^^^

SP3:U032WO(IQ

denotes -dn ^cbitraEy partia'l rjr.e^ within a p.icture, and a sizG; a sliape, a foat^rc, and Che like thcrenf ai:e nut ljrn-il.eif. Wamelv^ tti(? "block.'' includes, Tor example, the TU, the Py, the IJUU, the CU, the LCU, (.he .■oubblock; the 5 miHToblock, or an arbitrary area (prmr-eKs unit) such as a .^iice, Jn Piddition, otiior pavtiiil areas {proccs.'j unit} are al RO included. In the ca^e where the ^i^e, l.he proce!3rf unit, or the 1 "i ke needs to be lim> ted, appropriate deHi:ription thereof wi 1 1 be made. 10 [OO.iyj

Hijweuer, in oi-der I.e achieve a hiqher crtcad Lny efficiency in the AVC ur HEVC encodinq schi^ine, it. la important to JieleoL an appropriate pri-'.di<:l.ion mode. IF, tOU40]
/^H ^n example of the afjyoc.i rtted selection scheme, there .i 5 ^ met.hod incorporated > n reference sottvjarG (d.i .'3f:S o.'^ed in
hl.i.p: //iphome .hhi , de/j-jiiehriny/ttnl/index .html of 20 H.^fj-i/MPEG-^AVC, v;h 1 rh 1 .s ^.:alled JM (Joint Model), [0041]
In t-he JM, two mode dctcrmi ni nij methods ot high co^fiplexi ty rmnfe and iotj complon ty mode de^3cribed belovj can be ^ielei;Led. In both of the mel.hinls, r.ot^L Lunction 25 valuefj wich respect to the reyper:L.lve prediction mode
modeH are calculated, and L.he prediction mode vjhore the
cohit function value i a rifinimized is selected as an
optimal mode for an ri.'SHOciated block or macroblock,
[0042]
.10 The CO Hi. fifnction in the hiqh compl e^i ty mrjde i^
expressed by the following FormuJ a [U-

Sf:iblQi2WO00

[W^L.hematic^l Kormiil rf 11
Co3t(f^ode e Q) - D I >.*!! ... (1)
[00441
5 Here J Q de.rioLe?, a total sol cL i:andidate mode 5 ^or
encoding the asisoci ^l.t^d block, or tnacrobliick, and D denotes di f fererice energy between a decoded imaqc and ^ri
input irtiaqe i n L.lie case of cneocli rig the associated predjctiosi mode. ?.^ denotes a Laqranqe mull.ipiier given in as a funcl-ion ul a quantization parameter. R denotes a total i:ode aniount i net uding an orthoqonal tr^.n^rorM coeTf icient in tVif^ j-^se o£ encoding tho iiMsociatied inodo.
Wamel y, v/hen encoding ].n a H i.yh compleKity -mode > 5 1^ intended l.o be pertoriiiod, "i n order (".o calculate the
above-dehicribed p^ramcter.-i L) ^nd R, a prolirainai-y
encodi nq process needs to be perloniied once in every
ca mil date mode, so l-b^L. t± higher calculation ariiuinit is
required. 20 [CHHb]
Tiie ( <■>-■-■< Iiinction iri the low corciplexity mode is
esprcs-sed by the tollowinq KorriiLil^ (2) .
[0O47J
[MatheiTi'3k-ioai Formula 2]
25 Coal:[^tode e Q) - n + QP2Quant (QP)'■llcaderhii l. ... (2)
[004R]
Here J unli ke I.he case ot the hiqh compl exj j.y imjde,
t) becomes difference energy between a predlcS.ed image and
an input imaqe. QP2Quant (QP) is given ^ii^i ^ fniiction o£ ■^n a quantizal-i oJi parameter QF, lioadcv-!ii |. i H rf code amount
with rcypecC to intocmation .i.ncl uded in Header j which

IV

does not liu:Iude the on.hogonal trariMform coef f ic.i f?nl ^tid
is ccillt^d a motion vector or a mode.
[0049]
Nsmcly, in the lov rrijinpiexity niodo, iilLliough the b prediction process in each candidate mode needs to be perf ovrniedj since a decoded iiitagc .i -■3 riuL needed, th'^ oneoding process needs not be pc?rn>rfiied. Therefore?, it is po.s.sible to "i rupleiiient with a 1 ovier ealculati on iiiixount l.h^n that .in (.he high complE^xH.y modo. ]0 {OdbOl

In the AVC, intra 4x4 prediction^ inLra S x ii prediction, nnd intra 16 x If, prediction cxi Ht. In the IIKVC, as illustrated in Fig. ?., iariynlar prediction J ^3
l!i applied to ^ x 4 to ft4 x bfi pixel bJ ock.'s.
tJanielVj in the AVC, a^ i 1 ] nr^trated in A of Fig. 2, the intr^ predietioii process is performed by 3-directiosi + DC prediction. I it contrast, in the HEVC. as
20 illustrated in ii of Vig. 2, the intrti ptedicLion procoj-iJ-^ i r. perloriLLcd by .i2-directior. -I t)C prediction. Accordingly, the prediction accurrfcy is improved, \Q(ib2]
In addi < ion, in the llJ'iVC, ^^ illustrated in h'ig. i,
25 planar prediction is defined, [0053]
III the planar prediction process, prediction pixel's included in the currc^nL block arc generated from neighboring pixels (pixels that are already encoded) of a
30 current block whic}i is a target of Ihe process by bi-
1 i near Interpol ^ L ion. The planar prE'diction process can

^P3ril032WOG0
iniproue encoding ef f "i ci t^ricy lor an area where there is qrnd^tion.
In I.he HEVC, as illustrated in Fig. A. the encoding h proccsjj i.« pt^rlormed in the intra prediction mode using three most probable rnndeK. Namely, an intra predi cLion mode {TJnovc) of n>
25 I'iy. 5 is a diagram describing MDIf, [Mode Dependent
Intra .■Smoothing) regulated in l.he HEVC.
[ncibvj
In the case oT I he AVC, as illustrated in Ki g, '5, a [1211/4 filteri ng process is performed on nei ghbor i ng 30 piKel^] of a current block in an intra a x 13 prediction iflodo. On Llie other hand, in the lll'IVC, oi\/l ieif. In 10 the ca.'ie where the block ^i i ^e ol the current bl or:k IH 16 X 16, in s prediction mode of a direction other than 3 directions cl o^'.e Lo the horizontal di recLlon and 3 direction!^ cTuse to the vertical ilIrffcLionH the filteriny process \:-\ applied. In the ca.'se ijhere the bloc); si^-e ol
t!j L.he current bl(n:k is 32 w 32, ?n a prediction ntodo of a di rei^l.ion other than Lhe horizontal directi "iis -ind the vert i fiial direction, the EiH.ering process i s applied.
FurVheniiore, in the IINVC, Toi.- l.he purpose of 20 reducing bl(>i:k distortion in the case where the
predici ion mode is the IX: rrmdi'-, the horizontcfl mnde, ;)i-the verl.l::(ii inodc, a bouridj^iy v^iue sinoothinq process ill u.'il rat.ed in Fiq. 6 i K reyiilaLed.
[oo^nt
25 For csampl e, In \\\^. case of the prcdi c» i on (DC
prediction) where l.he prediction mode i^ l.he DC mode, the tiltcnng process j smooth in q prrjceM^i) illustrated in Fiq, 6 is perT,jrmed on both of the neighboring pixels adiacent to the upper side (Top) of l.he i:ijrrent block which is a
30 tav-ge(. of the process and I }ie rieighboring pixo] s adjacent l.£) Lhe lett side (Lefl.) nT l.he current block. I n

addition, i n [.lie izii.^ift of tti'e prediction (hori voriL.Hl predict! on) i-;ht^re the prediction mode i -•■^ \.-hf iujriztjnl.al mode, I he Eiltering process [smooUung process) i 11 iiri(".rated in Fiq. 6 i ^■3 perEorciied on the neighborinq 1j pi xelij adjacent to the uppf^r sictH (Top) ot the current block. In the c^st- rjE !.he prediction (vertical prediction) whecf l.ht* prediction mode is the verlie^l (node, the fillering process (sifloothinrf proce.sfi) illustr?il.t?ii in Fig. G is performed on l.he neighboring
10 pixels adj'Hietil. Lu the left side [Left) of L he i:Qrrent block, [GGGIJ
In the ca.se wliefe (.he image is encoded/decoded by the abovc-deseri !>ed inter prediction, in the obl.rj i ned
lb decoded im^gf, there is a portion where? 1 ht- change of lufTii noHj Ly, color, density or the- like IH planar as illust r^Led in the area 11 of I'iq - 7, and a strip shape of density irreqularity (HIJ -called bending) occurs, so that the phenoTrn^rujn where contour is seen may occur,
?0 {00621
'rhi^rei"i>re. a contour noise eoi.jnterTne;i;iuri^ pi occ:]-] disclosed in Non-Patent Document P wa.'J pfopu.'sed. By perforin i ng the contour noise coi.intprmeji^ju i^e pL■(J[:e^is d"i Hi-loried in Non-Patent noeumeriL ?, i.hf. bending occurring
?5 ill [.he area 11 of the decoded image is suppressed as
illustrated in l''"i q. 8, y-n \ hni. .■smooth gradation can be
obtained.
[0063]
■['he r:r)ril.(jut noise countermoasure process d"i sel o^ed
30 in Won''Pa!.ent Docufnent 2 is described more iri del.ail,
Kiq, <3 i.'i a diagrajvi illustratinq an e>:oiiTipl e of rj behaviiir

""I
^iE^:ibi032WO00
o± the conl.injr iioise countormcasure p|■.-()^^E?^3s. In the
contour iioiMf; counteriviGasurc prnrcy,^, firstly, a thre.sliold value dctcrrtii ni mj process expressed by the rc)llowinc( Formut ti.'j j3J and (■1) is performed by u.-i i ny 5 neighboring pixel jj oE ci current block J IJ>Jstr,"itf the threshold value
IS rirJiKsHOLD is set to be nKt'<[ Lo 9. By the threshold
Vfilue determinj ng prriceHM, features of the nci ghborlng p i xt^.ls o± the curr^riL blcn^k is determined, N^mol y^ i (. i^] determined wheLher or not the periphery of [.hn current block 'yi a portion where th(? <:h^rigfc? u£ luminosity^ color,
20 den'^i Ly^ ur Lhe like is pi tiita r ^n that the bending can r.riujr. In the case i^'h^'j-r S.h;; rosult of the thrci^hnl d v;^iue dctcrmi ni ny pY:(n:ns,s is true; that i s j in L}ie i^a^je tJhere it i s deLeriiiined that the periphery of fc.he current block is a pur Lion where the change of 1 vm i iio.siLyj color,
2b density, or the like is'planar .■JO Lh^L Lht^ bending can occur, instead of the filtering process described with reEt^rence to Fiq. 5, a bi.-linear interpolation process exE'i^e^.'^sed by the following Formulas (5] to {9] is perEormed on the ne i ghboring pixels ot the current b1ock
30 i I ln,'^trated m i''iq. 9,
rnofitij

^^

St>:^bl03SKO00

[Mal-fiemat.i.cal Forrtiu] a 4]
rcf LO] - refLOt , . , (b)
reE' [ij - HL -I- i* {Al. BL + Ml /7W (i - 1 to 2M -
1) ... (6)
5 ret' {?[^1 = ref [?N] , . . (71
rof r^N -I- i] = AL + i^ (AU - AL + N)/2H (j - 1 to 2N
I) ,.. (R)
ref (4N] - ref [4NJ ... [9)
[nOGT]
IG Tlie profie^s is applied to only th^ 32 x :i2 blocks
and ci riag rcprc^jtntinq whf?Lher or not to app] y I he* prof^fss (on/orc] is rcqvil iiS.ed in a HFtqu^ncc '^^ frame ter set
fncita]
i !i lloveyer, in L.hft method disclosed i n Nun-Patent
Document 2j the threshold value was set to be fixer!. Therefore, there Wrfs a problem in that, in the sel £:r:[.i[)ii of [.he filter i:riy procesi-i oii the nei yhbocing pi xel -c. in the i iiLra pred i rtioii, appropricite sel ection is nol. performed^
20 .'id that .im;jye quality i a (n:bnece-'5:jarily deteri orated. 10069]
\x\. the ttireHhoid vat 0(? determine nrj process, a.'j exprcfjried in Vnrniulas (3} ^nd ['IJ , a val nes calculal'erJ from pixels uMlue of the rit^iyhborinrj pixels of the
2b (Current block and the threfihotd vciluf! are compared with
Oar^li other. However j i T the threshold value ^ ^- .'ie t to bo Fixed, .in Home cases, it is consiilered that, I.he value is rujt appropriate to, Tor example, the bit dcptti o£ the image data is consi dered. In t!i i i^ case, there i*, a
30 problem i n th^t, a.-j a result of the abovc-deHi;ribed throsholiE value deL.ETmininq pror:i=;iS, a proper

-..^'
SP35I03?WC>0CI
detcrrrii iirfLion result is not ahl.ained, and an approprial.e
filter i.'^ not selected^ ^o icirage quality of tho decoded
imayf! is unnecossan 1 y del.etrioratcd.
[nu70]
5 In add"i 1. i on, it is considered L.hatj tor example,
the bit dt^pi.h ot the imaqc dat^ i^ changed in intcrn;j.l calculation at the time of ernx)ding/decodinq. in Ui i .^ casc^ simJI^riy, there is a problem in that a f .i xed threshold value is not ^ippropriate to the bit de?pth in
10 the -internal calculatinn, rind an improper rc^iuls. i>r the l.hreshold value determining process "i .=■ obL^ined, so that iinaqe qusi i Ly "I the decoded irnraye is unnecessarily detcri or^l.ed. (0071}
15 In addition, the linage quality of the decoded iiitaqe
cannot be ad'iu^l-fdj tor example, by the user or the like iadjustinq the Uire.^hold value. [00721
There tore, i n l.he present technique^ Uhe l.hreshold
20 v^l ue is set to be vijir Liible. For example, tht^ I hte.sliold v^lue ii ullowed L •., be set accordjiiq 1.o L.he bit depth o* I he imaqe dat^ , By doing so, it "i s poHHible to suppress a deterioration i n im^^ge quality. In add! I. i on, by ^setting the thre^ihold value to be variiibl e, it is
?.^i possiible to ad'iif.st the image quality of the decoded iinaqe L0073]
Nf;x.L, \jith respect to I'.he Mbove-described present technique^ examples of cippl ir:MtJoii to specific device'3 wi 11 bE" ifeseribed.
30 [007^[
<1. First l';nibod iiiient>

SP351 03?wt)0n

Fiq- 10 is a block, diagram illustratinq a conf iqijrjiL.lcjji exavtiplc of pin image eticodinq d(?v i CE? as a kind a£ an image proceHsing apparatus to which the b prcsf^jiL. techniqtic "i r, ^ppiied. The imoge encoding do vice lOCi illustrated in Fig. iO encodt^.s image data of a moving image by u^>i ng ^ for ezamplcj a pi:edidtion pcoco's.'j of I.he HEVC or a pred i i^l.ion process in ;-3 Ht^lieme equivalent. Lo the IK^VC,
10 [00751
As illustrated i n Fig. 10^ the im^ye f?ni;oding device 100 i'3 configured to include -jn A/D converter 101, a 3C.roen rf?firi/angeiitent buffer 102, an arithmetic uni I 103, an orthoyotial transform unil. 104, a quantization iiriLL iDh,
lEi a 1oH^]les& encoding un i. L. 10&, an accimul at> on buffer 107, ^n inverse quanti i/raLicn unit I OG, ^nd ^ri inverse OTTl.hogonal transform unit 109. In addi t i on, the image encoding device 100 is confiqured Lo include ^n arithmel.j c unit 1.10, a loop filter 111, a frame memory
20 1.^2, an l.nl.ra prediction unil. ll.i, an inter prediction uniL ll''±, a prcdi ct.".."f u[i;igc selection unit '15, .^nd a rriL.e control unit IIG. fOOVf,!
The ft/D converl.er 101 A/D-convorts i npul image data
75 and supplier lihf converted iinaqo data (di gi '..al d;^La) to the screen rearrangement buffer 10? [.a store the converted irciage data. The screen fe^rtangement buffer 102 rc^<"ranges the iraagcs whi ch are in the stored frame ordGr for display by using the frrJiEie order for encodinq
30 according to a GOP [Croup of fLiil.ure) and supplies the imaqeK ot vjhich frame ordc^k- i^ rearranged to the

^^

:i3':t^ici32wooo

ar.i l:hiiipLiC unit 103. In addition, the screen rca rrcingeHient bk.)ffcr 102 also supplies Un? images of wh L[;h trartic order LH rearranged to l"he inLra prediction unit 113 cmcr I.he inter prediction uniL 114. ^ iOO"/?]
']'he arithrtiotie uni |. 10!i subtracts thie screen learranqemenl. biiEEer 1U2. [U07G}
The arL[iQgonal transforrti im I I. 104 performs orthoqontil 1 r^insform such as di scrt^I.e cUiiiie transform or Karhunen- [.cieve on the difference J nEormation supplied
25 from the arithmetic unit 103. The orthogonal transform uni t 104 supplies a tv-ign^'ifoon coefficient thereof to l.he qnancization um t 10!i. [OOVyj
The quant 1 yaLJ on unit 105 performs qu^nl.ifiril.ion on
30 the transform i:i>E?EEioient supplied from Uie orthogonal transform uni < 104 . The quantization un i L 105 sets a

SP351032WO00
quantization pariimoleL^ based on .i.nformal.itjn on a tarqct Val ue o± the code; rirnount suppli ed Troiii the rate control unit 116 and pf^rforms quantival ion thereof. The quantizcition unit 1U5 3upp1 i ed Lhe quantized trrtn^jlorm S coefficient l.o the los^le^is encoding unit lOfi-
[cosni
The I053I c-r-.'i encoding unit lOf^ encodes the tr^nstorin coef f i cLe^nl. quantized i n Llit^ quantization un i I. 10 El in an arbitrary en cod i nij .■schetvie. Since the
in eoeffici£-.nL. data arc qi.iiintiy.ed under the control ol the
rate nonttoi unit 1 IB, Llie code amount be (or the code amount is spproxiHiate to the t^rgel. value) . [OOGI]
lb In addition, the lo;isle&& encodinq unit 10(]
acq^i i rfts the inf orm.-al.ion representinr| i.he intra prediction mode or the like troni the mtrii prediction unit 113 and acquires inform^^tion representing an inter prediction mode, the difference nioLlon vector informal, ion,
20 or the like from Liie inter prcd.i cl. ion inii t 114 .
iooe2]
The .1 oKhltss cncodi ng iinil. 106 encodes the Vtir i OIJH kinds of i nformation in an ?i rbL trary encoding scheme to be used MH a portion of herjder intoriiiation of the enr:oded
25 data (Hometimes, referreil to as an encoded strej^cn) ^ The 1 OHH1€?&S encod.i nq unil 10b supplies the encoded data obtained through the encodinq to the acciiiiiulation buffer 107 to accuniul ^l.f? the encoded data. [O0G33
30 The eiH'oding scliciae of the I o.s.siej,s encoding unit
106 include.'^, tor example^ var i ^ble length encoding^

2^ ^
S\'3blQ?>?MO0{i
ari (.lirtietic; encodingj rirHt Lhe like. The vj.v\ abl t^ I eiigl.h ein:ode includc^r ^'-^'- eXcSfViple, CAVLC [Contexl- Aildpt.iue Variable l.engUli Coding) defined .in the U-IbA/AVC scheme and the like. The arithmetic eiicodtf includes^ for 5 example, CABAC [Context-Adapl.J ue Binary Arithinoti c Ciiding) and the like. E00S4]
The accumiil ^Lion buffer 107 terrporj^rlly stores the encoded dai^ supplied from tho lo^iHle.'s.^ encoding unit 106 10 Tho accuTiiiil^Lion buffer 107 oiitpiiL,s L.ht^ stored encoded
daLrt l-U ati outside of l.lie image encodinq dcvi r^o 100 al. a predetermined 1 irniiiy. Kamely^ the occurnnl al.ion butter 107 i5 al^o a transmitting unit whirti transmits the encoded d^l.a. 15 (OOe^l
III addition^ the L.r^nHfarHi coefficient quan t i v.ed in
the qu-ifiLization unit 105 is also supplied to the > n verse
([Uantization uni < lOE . The inverse quant "i ypiL: i on iiniL 10a
pert onn^ i nuer se quantization on the ijuanLized transform
?it Goeff icient i IL a tdeLhod correspond i ng Lo L!ie quantization
of the qii^iitization unit 105. 'i'he inverse qnantization
uni t lOH oupplies the obu^ i nt-d t.ransform coeff iciont to
the i iiv(?rf:,e orthogonal rrj^n^] furm unit 109 .
[00861
25 The inverse orthogonMl (.r^nsform unit 109 performs
i n verse orthogoritil l-r^iiHlorti: on the transform coef Ti r i en I. supplied from l.he inverse quantization unil. 10!! in a method cor re y pending to the orthogona'' t r ti n ^i ED iici process of the orilioyorial transform unit 104, The inverse¬ly 0 orthogonal-transformed output [ resl.cired difference
infonnrjLlon) is supplied to ihf ^iiis.hmetic unit 110.

SP:ihl032tJO00
(ooavj
The iiritbmeJ i.i: unit 110 adds l.ht^ predicted imact*^ supplied from Hit* Intra prodi ct"i on unit 113 or the i fi< er prediction unit 114 through Hie predicted -im^ige i^electiou h unit 115 to the recovered dlEIerence informal, i on i-^hich is thf. Inverse orthogorml transform resiiH. .'iupplied. from the inverse orthogonal transform uni 1. KSy to obtain a 1 ocoil ly recovered ira^ge [hereinafter, referred to as a reconstr^id.ed image). 'I'he reoonstructod limage i^
IG suppl 1 ed to the loop f i 1 l.fr 111 or the 7nt ra pr e?(lli:tion uni L. 11:J. [OGHliJ
TEie loop rilLftr 111 appropri tilet y performs a Hi 1terinq prore^H including a deblocklng fliter< an
15 adaptive loop filter, or the like on the reconslrructed
imaqc ■supplied from the ari l.hnieLlc uuit 110. !''or example< the loop filter 111 removoj-j block distortion of \.ht: rccoiisl.rueted image by pE?tIorming the debl ork i iig f i ll.ering proces.'-i on the reconstructed im.Hgi?. In
20 add]tionJ for example, the loop filter 111 improves the
iiiiage gual i l.y -ly pertorrtiiiig the 1 oop filtering process on the debl ocking tilterinq proce,'3s rt^sult (the reconstriioLed image from whir.h I.he block distortion J5 removed) by using a Wiener V'i 1 L.er.
2b [00&9S
In addition, the loop filter 111 may further periorm any other ;^rbitrary filtering proceHH on the t-econstructed Jtriagf. In additionj if nece.'^ya ry, the loop filter 111 Tiiay j-nipply inforination 5uch a-'j a Ellter
/iO coeiticient whj ch is used for the f i 1 tering process to the lossl er,s encoding unit 106, so that the information

^ >
.P3!Jin.l2WO00
may be encoded. [00901
The loop ri iter IJ. 1 5uppl les the f i.l teri ny process result (hcrei niif ter^ rcfrrned to as a decotfed iinago) tn 5 I.he Iramc nieiiiory li2-[0091]
The f roiTiP meiiiocy J. ( 2 .'3[ (ifes the suppl i t^d decoded imiige, and at a predetetniinod L.imingj the f(-;iiiie memory supplies tht^ s Lured decoded image as a re Terence imacfc t" 10 the inter predlcl.ion unit 114. [0092]
The intra prcdicLion unit \\3 performs intrii
. pred-i cL.ion [predict Lon \jithin n screen) ot qcner^^.ing the
pred Lifted irtiaqe by using pixel values in the process
15 target picture which is the? reconstructed iinrjge supplied
M.'s the referein:e iKiage from llie arithmet i c unit 110. The
■i riL.na predict"i on unit 11 3 per fcriEks the intris prediction
in a plurjilil.y o"£ predetermined intrci prediction modes,
[00931
2Q 'I'lie intra pred i cl.iun unit I 1 3 rjenerates the
predicLed imai^O--j i-n all the intra predict ion modes which ^re CandidatcFi and evaluater> l.he cu^t tuncti nn values ot \ }ie respective predicted ImageFi by using the "i ripuL image supplied from llie screen rctirr^iigement buffer 102 to 2b select the tjpl.imal mode. <:)jn:e the intra preili ction unit 113 sc] eel.s the optimal i riS.rti predict.i on [nude, the intra prcdi cl-ion unit suppl ie;s the predicted image generated ' n the opt- Lriial mode to I be predicted irTitiqe selection unit 1 ib. 30 [009^J
In iiddi Lion, as descri bed abovej the .intra

prcH i (:!.ion unit 1] 3 appropri iitcl y .supplies intra prediction mode infoirmatinn re^pre;senting the selected intra prod let i nn mode or Lhe like to the lo^ijlcs;? encoding un il. 10 &, so tliat encodinq i s pe^rEorri^ed. El [0095}
'['he inter prediction unit tl4 pE^rforms an inter prodif:l-ion process [inotion pred i (:(.ii)ri process and motion compe.risaLlon. process} by i-i^^i i1 *-l'^ input image supplied from the screen rer-irrMiigeiiient butter 102 and the
10 i^elerence imaqe suppl i ed from the Ir^iino memory I \? . Mnre; specif ical 1 y, l.he inter prediction umi t I I'l perTdrin;s the motion compensation process according [.[) Lhe motion ueel-or detected by pcrfornii ny Lhe mol.ion prediction as ».he inter prodict.i on prnc;f?HH L.o generate the predicted
15 image [Inter predii:led image intormation) . The inter prediction ufi i L 11^ performs the inter predi etion i n a plurality of predeLermined inter prediction mode:^-[00961
The inter prod.i ct i on uni L. 114 yenerates the
20 pveii Li:Led images in till I"he i iii.er pri^diction modes which fii-E" candidates. '"'he tn< er pn;di(:tion unit 114 cvaluc^tcc 1 h(? cost function v:-jUje cjf each predicted imago by usinq 1 he input imaqe -^uppl i ed from the scj:een rearrangement buffer 102; the i n rriMiial ion of the generated difference
?.h motion vector, aud Lhe like to select the optimal mode-Once the inter prediction unit 1 M 5elect.'3 the opticiiiil inUer prediction mode, the in^t^r predi.i^Llon unit supplies Lhe predicted imago goner;jLed in Lfie optimal mode to the predicted im.aqc sol e(;L i on unlL lib .
:in [009V]
The inter prediction unit lid supplies the

^^^'

.^P:3bl032WO00

informal ion rttpresentinq the scl(?r:L.t^d inter prediction mode or 1 he intoriit^tion ncccs.'j^ry Eor pertorGiinq the pv-o'ies.'i in the int<:r prcfiicl.Ion iiiode at the time of f[ei:oding the encodt^d dril.a to the los^leG!? en cod iny unit 5 IQ*.!, so that the encoding is perfornicid. 'I'hc fK^oes.sdry
intormat.i on i rid udes, tor example, thi-; i nfonii.itioii ot the generated dillerence motion vect-or, rj flag representing an i.ndex uT. as ttie prediction mot. Li>n vet:tor as the predioLion motion vector i rinjnciriS.ionj and the like.
10 [0093]
The prcd.i eted imj^ge .^election unit 1 15 selecl-.'i (.lie supply source of l.he predicted irnaqo which IM (.O hi-supplied L-o Lht^ arithmetic unit tO.1 or [.he arithmetic uni t 110. For example^ in the ai.^f. of the intra encodinq,
1 Jj the predicted imaqe sel ecL-i on unit 115 selects the .int r^ prt^diction unit 113 ^p-i L.lie supplying source of the [j;:t^dieted imaqe and j-juppl J ft^^^ the predicted image supplied trom the intcri prer.iiction unit 113 to the ari thciiE^I. i i: iinil. 1 03 or I.he arithmetic unit 1 ■) 0 , In radd iL.ioii, Eor example,
20 in the ca.^f ol the inter oneodiiiq, the prer] i cl.t^d image
pjele^cLion unit 11.5 select-^ l.he iiili^i: prediction unit 111 ijH l.he supplyinq source n-yf l.he predicted image and supplies the predicted icriage supplied trom the inter prediction uni t 11^ l.o the arithmetic unit 103 or the
2b arithmetic unil. lln. (00991
Thf^ rate control unit I 16 control H a i-r3te of quanSization operation of the qu^nl i L^aL.i on unit iCIh ba.'ied on l.he code amount of the encoded d^l a inx:umulated in the
30 accumulation buffer 107 y-n thiil. overilow or undertlov; does not occur,

^^^

;jp3:uoi?L^oon

[OlOO]
The "iraiiqe? eriixjding device 100 IG confi glided Li) further include a threshold val uc r-f.y.i. Ltig unit 121 and a f i t tei.-lng processing uni t 172. ^ [0101]
The thrcGhol r3 vai lift setting unit 121 Gcts l"-he threshold val uianipl e, (.he threshold value oett i ny unit 121
1^ may set the Hireshuld value according V.n the bit depth ot the iniisqi^ data which arc the cneodi ng L^ryet. [0103]
In this casCj for exarpp! e, the threshold value .■letting unit 121 i^ny in:qiiire the information on the Vi i \.
70 depth of the image dal a wliich are parameters tr^nsmi {.\ ed as the scquericj' par.itrtetcr set [SP^J or the like Iroiii the lossless encoding unit 106 to the der-rid Lrig side^ determine the bit depth of the imaqi-. datri based on the informal- Liin on the bit depth, and wet 1 he threshold value
25 accord lug to the bit depth. I n addition, for example, the threshold value se( ting unit 121 m^y acquire the image informat i on [image data or information on i.ht- iMrjge data) from I. he ,'3i:ieen rearrangement buffer 107, detE^miJ ne the bit depl-h cT the image data based on the i rri^qc^
30 information [by atialyiiing the image inf ormati on) , and set the threshold va I ue ^iucording to the bit depth.

;js^:5.si():i2wooo
[0104]
III addition, i ii this ca^c, Tor examplej thf threHhoid value .■^el.Ling unit 121 fiiay tipdate L.tie threshoJ d vraliie by bi t-shiiting tho initial valup (Tor example, 8) 5 of rt prodcl: en mined thrcshol d ^alue accov-ding to the
determined bit dcptrh. At this time, for example, a value appropt-iate for \ lie^ case where l.hf? bit depth \ y- B bits lEay iie predotermi rit?d as the .i n i I ial value in fnivanee, and the threshold value setti ny unit 121 may hil.-shitt the
10 n ni l.ial -VD.} ue Jiocotdinq to n dittereneo i n (.he iiumbGr of l>it3 between an aetual bi 1. depth of l.hc^ image data and the 8 bil.s. [010^]
In add.ition, in this c^^ie, lor example, L.he
15 tVii^eshold veil ne setting uni L 121 may .'jupp! y the tlireshol d ualue inforiEiation jrepreHE^nting the afl.ef-UE>dating (.hreshold j-iiue to the 1 D^s-siess eneodi ny unit 106, so that the sequence par.aineter set {yP^>) , the picture pay-^ineter set (I'T.S) or the like fiiay be transmi l.i.ed to the
20 decoding side. AL. i-.his time, for example, the L.hne.shold v;-31 ue sotti nrj iiniL 121 m,.iv t?iLL;ode (for cxr-iinpJ e, golorab encode) the threshold v^lue intormati on rami supply as l.he threshol d value oiicodi ng information . In addition, for exaitip] e, in the case where tlie threshold v^lue sett i ng
?h unit t21 determines (.he bit depth oT l.lie image dat^ based on the image inf<>:^mation, and l.he threshold! ual ue .■setting uni L. supplier l.he information un Lhe bit dcplh LD the 1 o.-i;sless eneodi ng unit 106, ;") that the sequence parameter sel. (^PS), the picture parajfieter yel. (ITS), or
30 lhe like piay be transmitted to the deeodi n^j side. At
L.his tirtie, fiur example, lhe threshold valuta setting unit

sp."ibio:i2wooo
1 ?1 [jmy ettcodo (for ex^icnpie, golonib-cvicodo) the
■i iirocmation on the bJ L depth and supply ^H L.ht^ bit depth
encodinq 3 nforrnaL ion.
[01061
.■j In rjcJdil.ion, in thi'S c^^e-- Eur example, the
thrcjiuild value setting isriil. 121 inay generate flng infrju-matiDsi (thceshol^l viilue change flaq) repre.seitting i-jht^t.her o£ not the thre.'^hold value is updtit ed (changed) and supply a L.hreHhold value change EJ ag l.o the lossless
10 encodinq unit lOfi to transmit the threshold value change flag tn l.hf? decodinq side. By iloing so, in the decoding srde [Eor example^ the iiiiaye decoding device), I I. i.^ poH^sible to easily identity based on the vtiliie oE the (.hreshold value cSi^nye tlag whether or noL. (.he threshold
1^ value is vipdaL.F?d (changed). Namely, in the decoding s.i de [for exfim^le, the iiBage decoding device), it is possible to easily control whether or rH>L. l.o perform the process of updating [ changing) l.hf threshold value si mi 1 .-irl y to I.lie encoding si de [Eor example j the image encoding device
70 100). [0107)
In .^ddlLion, for example, the L.hre.shold value setting unit 121 may set the thre^][iold value according to externa [ detsignation such os u.^t-.v ' s dej,ignation. In th.i s
25 case, the value designated by l.he (►.■set or the like
eorre.'^ponds to the above J (irrLl-)--(?ricode) the threshold value information and supply as the threHhold value enct>dinq .i nf ormation. [0109]
10 Jd y^l ue .■^el < i ng um t 121 Ttiay generate a Lhre.shol d value ehange flag and Kiippl y L.h(^ threshold value change flag to the lossless encoding \tn i L. 106 to transridt the threshold value change flag to the dt^c:raTiipl c, the -threshold value setting unil. 121 iiirjy update (change) the threshtjld v^l vf extornally designated by a u.'sei or the T ike according to I hr bit depth of the image data vjhicli IH l.hi^ rncod.ing
?Q 1 ;^ rye I..
LUlllJ .
In this i:^y,f, ^ar example, the threshold val iii? setting unit 121 may ^crjui re the information on L.he bi t depth of the image lia \.?i Troni the lossless enoodi rig uni t
2^ 106 and determine the l.)j L. depth of the image data bMsed on the information on thf? bi L depth. In addition, ioi example, the threshold value Hetl.inq unit 121 may acquire the i.maqe information from the hicret^n rearrangement buffer 1 02 and determine the bit depth of the image data
."50 b^Ht^d on (.he .image information. [0112]

4ir^^

SP.TilOi^zWOOO

In addition^ i n l.hi^ case^ for cxampl e^ Lhf? tiirt^shold value sriU.Liig unit 121 may updiiL.e the threshold walue by bi I: ■shllLifig the thrci^hnld value externally desiqnateci by a user or the like according to the b detertni ried bit depth, at thiw [.iinen tor example^ the th^e^shl:)ld value appropi^i ate to the case where Liit^ hit dE?pth is S bits m^y be designatedj and the Lhre^hold v^lue settinq unil 121 may bit-shift the dE'^ignated threshold value aecordinq to ^ diCEecence in the number
1(1 ot bits bf^» i-ieen an actual bit dt^pLh oI the image data and e bi1H. [DllliJ
For exampl e, Lhe threshold val ur e.xi.er'iially designated by a user or the like i^ defined as
lb contour! riij artetact_thrcsholtl, The
contenting artefact threybiild is designated as the v^il ue cor re.'^ponding to the case where the bit depth .of the i][inge datd i-ih? ch i r- l.lie encoding target is 8 b i I.H . In the case sphere Lhe actual bit depth tjf l.he image data is
ZO n bits [n ^ R), the threshold value J-^el-l ing unit 121 bit-shifts Lhf: .-:i:jntouring_artefaet_L hre:ih()ld by using the f ol 1 OVJI ng Formula ( t 0) . [01 Mi [Mattieiiiritical Formula 5]
2b conLi)uring_artcfact_threFihoid « (n - Hf ... (10)
[dllbj
In addi t i ocij in this casej for ezampJ e, lhe threshold viil ue He LLiny unit 121 may suppl y the LhreHhul il value inf orriiaL Ion representing the af tcr-updal.i ng
!iO threshold y^lue to the lossless encoding un i L 1D6, HO that the ^equt^noe parameter set (yp.'j) j \. he pi cLure

SP31it037i']O(IO
parameL.E^r sot ), or the "like may be tranGmi t<-ed to the deciding side. At (.his time, for
lb example, the threshold value ^et* i ng unit J2I may eiu:cde [for example^ qol imib-encode) the threshold value informs Lion and J3iJ[>piy as the thre.shmd value eneoiliny i nloriBation -[G117S
20 ?^t this Lime, in -jidditiun, the thrc?jh(j]d value
sotti ng Uiiit 1 2') jf,r,y .-iupply the i n f e r-[(ia Lion on the bi t depth Lo the ] o^y-^ e.ss encoding ufi i L 1CI&, so that the segueEH:e parameter ^et (SPG) , the |>i cLure parameter set (['P.'^)< or the 1 1 !ie may be transmiMeii to the decoding
25 Hide. At th"i y Lime< tor exsmpl e, the threshold vtil ife
Hettiiig unit 121 may encode [for example; qolomb enoiide) the informsLiors on the bit depth and supply ae Hie bit depth encoding information. (01131
iO In addition, !n this case, for example, the
threshold valuci sel-l.i fig unit 121 gcncrti Les the threshold

value ohange tlag and suppl i-c..^ the threshold val ue nVir-inye f 1 ag Lo the lossless encoil i ny unit 106, so that the sequence parameter M^L {r>PS) , the picture p^ r^mel.eE set (PPS}, or the t i !;e iiirfy be transmitted I.D Mii? decoding 5 .■iide. [011^]
In addition J l.tie l.hteshold valiie sett "i ng ifni 1. 121 rp^y perform the self i ny (updating) of the thre^iho.! d value based on flrbi l.r^^cy parameters othe.r l.lian the bit depth,
10 In addition, although tho according (.o initial value of the threshold value is arbilrrtry, the initial vaJ ue rntiy bo, n>r ej^amplG, "8". in ?iddlLion, by sotting "0" i-is I he Lhre&hold vihlue, Hie l lireshold value setting i]ji|l. 121 prohibits llie bl-linear intcrpo.l at. i on pi:(K:er^s troin being
lb applied, y-n lliat the tiltcrJng p^ore^3^i described with
reference l.o Fig. 5 can bo applied. Namely, in the cjse whore the threshold valuo is "0", Hie method disclosed in Non-Ca<.en|. Document 2 is di.s?ibli?d. [01701
20 'ftie iiiltering procesfj i ng unit 122 acquires the
threshold valu." i nliirmation trovf, the threshold ^/alue setting uni I 121 and perform.s the fi 1 I e»- Ing process on the neighbor i ny pixels of the current bJ ricSi which is a target of I.he intra prediction prorei^iH by using the
25 threshold value. For example^ t.he TllLering processing unit 122 (tiay perform the threshold u^lue determining pror.c'iy. expressed by Kormul rjs i'-^) and [1) by using the |-.hrt^shold value iicqui red iLnjiii the threshold value sot t> ng unit 121 to 1 douL.i f;;/ features o£ the neighbor! ng p i xeLs
.10 of the current bluck. t0121]

"^ £P3!3t037WOGG
"tci addlt? nri, lor cj^ampls, in the r.at^s where l-lit^ determination re.'^ult I5 fial.se, that i H- In the ca,'3e vjiiere th nl.r^ predict] on if nil. 113 performs 1 nl 1.-^ predict 1 on by using the ^afLer-filtering-process neighboring pixels. By doinq y-n, the intra
30 predict iim unit 113 may generate ii predieted image reflect Lng a rcsul( of the fliter 1ng process,

SP35l03?VJOaO
[0124]
N^Tiicly, as doGcri bed abovo, i^i nct^ the thror^hold value sctt-ing unit 121 czan set the L hreshold v^alue lor identi fylng the fp-iL.Lires o± the neiyhborinq pixels o± the ^ curreni- block in L.he intra predi.i:Lion proeeM^s .at the time of ftncodnnq the irtiaqc diil.'i accotdn ng L.o the bit depl.li of l.he imaqe data or the like, the i nl.n-a predict! on unit 113 1-an goneriiLe the predii:l.ed imaqi; rellectinq the fe.suit of the f' ^ (.erinq pronesri appropri -ate to the iTn^yE? data.
to Namely, Lhe iraaqe encoding dcvii-e iOO can suppi^ftss
occurrence of noi se such as bending in the decoded iinaqe and carl suppro.'j.'i a deteriorat i on in irtiaqe quality of thc! decoded iiviaqe , £Q12b]
1 !j Jn addition, ^y. described tiboue, since the
threshold value seM.lng unit 121 can set the Lhre^hold value tici-ordinq to exl ernal do'^i qn^l.ion such ay user's dcj-i qnation, the Intra prediel ion unit I 13 i:an reflect the? external def^ignation ?uch as the mer' H designate on
2U of tlie image qu:al iLy on the predicted image, Namely, the Jmage OncodLug device 'LO(! can control I.he imaqe qufilily ot the rieiToded imaqe, [01261
I n addition, nr- described wi Hi reference lo Fig. 5,
2h Formulae (5) to (9), and the fike, the laethod where the filtering processi lock size oi7 (.he current bl oi:k) .
30 S0127]
1n addition, in Lhe case where Lhe bit depth oT Lhe

SP3510:S2WO00
Linage data is as sifiall as, tor example< U bits, the bending [contour distortion) illustrated in Fiq, 7 is remarkably ob5erv(?dd [[nwnvf?r, in lilie npu-if? vjheVe 1 h(? bit-depth \ ?i nr, 1 rJMje a^, for example, ID bil..-], Lhe beiidi [ig 5 -i '3 supprE^Mj-ied [.i fs rioL. v LHii^Ily iM)(i.spii:iK)ii.'3) . 'I'here fin e, an upper limit o± the bit depth vjith which the bi-linear tiltering process disclosed in Non-Patent Document 2 is applied may be provided. For example, the bi-linear tilterinq process di sclo'^cd "i n Non-Patent Document 2 may
10 be applied to only the caise where the bit depth is 8 bits, and the b'--linf-!ar filL-ering prcirTe.sj-i rrii-jy JKIL- be; iipplit^.d le the othe^ r caHes . [ni2J{J
In addition, the bi-linear filtering process
15 disclosed in Non-Patent Docvxmont 2 may bo applied to only the processing of br.i qhtness s.i.qna.l s. l[owever, the bi¬linear filtering process may be applied to the process of eol o^ rS L EEefence Hignal H . Namely, Lhe ^ir^efsenl. t.f?ehn i que jfiay be applie<[ Lcj [Milor di rEereiii:e sjgrial.'s ay: v;el 1 a^s
?0 bL-ig^lLlle^3a Hi glials . [0120]
In addi t ion, i n the case vfherc the input signal is 4;^:4 or R
fig. ii 15 a block diagram illustrating a main eonf iynration o^arnpl e-. at the thrc^hol d value setting mi i \. 171 and the "Ti 1 (.ering processi ny unit 122 in the <:?iHe
JO ^hi?re the thre^.'^huld value -^etl-j ng unit 121 of l'\q. IG ftets the threshold value isricording to the t>i(. depth of the image ctata which > s a target of the encoding. {01371
In the example ot Fig. II, L.}ie threshold value
1 5 ^.(^tting uni t 121 is eonf iqured (.o include a bi t tJep^.h
determining unit 131, a thn-shold value bit shirting unit
1J2 J and nn encoding um t 1.13.
[01331
Thet bit depth di-:termining unit S 31 determines the
20 bit depth ot the imaqn ilal.rf as the eijcoding I arget and supplies the i n rcjuniLjtion reprcseni i ng the bit depth to the threi^hold value bit shifting unit 132. [0134]
Kor e:^ample, the bi t rSepI h determining uni t 131
25 acquires the informafion on the bit depth of the image data from the }os.'-n ess encoding un.i t 10(i and determines the bit depth of <-he image data basei^ on the intormation on the bit depth. In addition^ for t-^x^rciplet, the bit depth detE^riiiining unit 131 acqu i res image intormation
30 troiR the screen rearrangement biiftet 102 and determines the bit depth ol the image dat^ based on the image

GP35(032WOnn

inforin^l- Lon. [01351
Eti addition, in the. r^t^e of transrtiittinq l-he n nfoiiiiation roprc3fn Lhe bit depth dctcrmi ning unit 131 ,
15 [Ol37t
For example, i n Llie l.lireshold value bi t ;>lii fLlng ufi LL li2, the vaJut; appropriate to the case where Lhe biL depth is 8 bi LH is :^et as the initi ol v?il UE^ OL the thrcsbol d v,^l iie in advance. Once the ] hrE?siLuld value bit
20 shift in<5 uii i L l'.V2 diL;quires the m fornix Lion representing tiif- ■>iL depth from the bLl. di:pLh detcrminitiq unit 133, \..\te threshold valii^ bit shitting unit bit-shi fl .■■^ I hf i nitinsl value by ^ if i rCerence in the number of bi ty between the bi t depi.h of the image data represcnl ed by
2h the information and the C bits. [013S]
The Lhi-eHhoid value bit shifting imiL 1:^2 supplies the i n forinaLion [ threshoJ.d val ue i n foMiirfLlonJ representing the after--upd^Li ng {ohanging) threshold
30 vMiue to the neiqhbori ng pJ xei determining unit 142 of Lhe filtering procey^i i ng uni L 122 .

.6.^^

yp3'iin:52wooo

[013[S]
In iiciili I- Lon. in the case of transmi tt i ruj L.[ie thrc'^ho! d y^iiie intoriiiatioii representing I he rffter-updstlng (changing) threshold v^alue L.ij the decoding side^ 5 the Uire.'shoid value bit shifting unit i:^2 also supplies the threshoJd val ue Lnlormation to the encod.i ng uni l 133. [0140]
The oni.'^;^iless encodi ng un L L. IDG to transmit the qol onib i-ode Lo the decod i ng jiidc. [OMl]
I ri addition, for e^^ample. In the case of transFELL tting the threshold u^.l ue information reprcscnti nq
25 the a£ ter-updatini (ch,^ng i-rig) threshold value to (."he decoding sidc^ the emxiding unit 133 acquires the threshold value i n ror/rci^ tion representing the after updating {changing) (:hre.sh(iid value from the thro'^hol d value bit shi ft i ng unit 132 and golomb-encodes I ht;
30 threshold val ue i n format ion. The encodi nq un i I '■133
supplies the obtrained golomb code of the Lhre.'3hi>ld value

informsl-.i on [sometimes, refer re(f Lo aa threshold val ne
encod> ng intoritiation) to I he lossless eiicodi nq un i < ICIlJ
to l.r^nsmit the qol ijj^ib code to the decodi ny Hide.
S(I1^2J
5 In addition, (.he eiicodinq un.it 115.3 rnay generate a
threshold value chi^nqe f ] aq rep re? renting whether or noL. the threshold valup i j? updated {changed) and supply (.he threshold value chiinrfe f 1 ^ly Lo the lossless enrod i ng unit lUi; to transifli t »he? L hreshold value change Elag to the
1 0 dei:uding side. (01431
I n r-3ddition< as illustrated J n Fig. 11, the fnlterjiig processing unit 122 i^ contigured to include a mode/block, size buTTer 141, a neighboring pixel
3 b determining un it 1<2, a filter dcci si on uci L L 1^!^, a lovj pass filter unJ L lid, and a bi-l.inear rlJL.er unit 145. [0144]
The mode/block s J ?c bu rTfr. 141 acquires the i n Torffirftion. (itiodc/b''oek f^iye) on the block size and the
20 mode of the current block v;j i h respect to the predicts on iiLodes v/h"' ch ?ire candidates troni the i r.tra pri'dicl.lon unit 113 and sL.oreri the intortviation. [01451
AL a predetermined ti.mi nr[ nr liaweif on ati e:rternal
2b reqiie.'s t, the mode/block si vf^ Vnirfer 141 supplies the
i:\ oted intormation (bl ock -ri i xe) on the block size to the neighboring pixel determ i ii i.ng unit 142 . In addition, at a predetermined timiny or b^sed on an external request, the mode/bl oek s i /.f? butter 141 supplies the stored
JO inf ormat i on (mode) on the mode and the stored i nroiiEEi.HL J on [block si^e) on the block size to the filter deeiKion

Mil i L. I-IJ . [nl'16]
The neighboring pixel determijiing unit 142 acqui ro;?
the neiqhbon nq pi xelH adjauent to the upper side and the ?> lett side of tho currenl. biot:k with respect to the prcd"ir< ioEL modes which are caiididatc"; from | he inLra prediction unit ] ] 3 . In addi I. i on, Lhe neighboring pixel deL.eriiiining unit ] 42 acqui ro^ I hf. (.hi-eHliold value intormation from l.ht^ Llirft^hold value bit shifting unit
10 132. In addil-.ion. Die neighboring pixel dctcrmimng unil 142 acqu i i/eM t.he intorination (block ■;? 7ol on l.he V>1 m-.k. si 70 Erom the mode/block size buffer 141 , [01^7J
In the ca.'if! or l.he mode where the current block has
15 a predeteriLiincd sJ zo (for example, lil? s 32) (or a size wi Vhi n a predetermiiied range) ^ the nf^i ghboci ny pixel dote rill J ning unit 142 performs the threshold vMlue dei.ermining process for j-iel ecl.lny Lht^ lilter tjhich is to be used for the f i 1 I.eri ng process on the neighboring
20 pixels acquired f roiii tho i ntr H pi ed.i ui.ion iiniS. 113 based on the ir.forrn.-jl.ioii on t!sG block size acquired from the mode/block ^3ize buffer 141 by using the threshold value informal.ion acquired from the threshold value bit shiff-irig iinlL l:i2. Natnely, the neighboring pixel
25 de.I.eL-EiLining unit 142 determines features of i.he
nfj ghboring pixels (for exarnpl e, cieS e^rciil rie.s whel.lier or not to be pixels of a porti on where the change of luiviinosity^ color, den si ty ^ or the like i.s planar) . [01'18}
30 The neighboring pixel determining unit 142 supplies
the detorrpi nation result to the filter decision unit 143.

-i

SP35i032WO00
In addition, the neighboring pixel determining unit l'T2 supplies the neighboring pisels of the current block acquired from the >riti"r i. rig p i.KE^l df-l.e refilling unit 142. The filter dei:iHiun unit 143 decides the type of the to-be-executed filtering process and the cipplicatinn method thereof by using the acguired
}.b inf o vrni-i I i <>ri iinil n eHu 1 L.. P''r)r exj^[ii[>l e ^ I he fi 1 I. e c deci si on vn] { 14 3 (leS.emilMe tJhelJier or no I. aciy of l.he l^ovi p^;;^; f I 1 Leriny process and the bi-iinertr liiL.erliig pint-.fy.y, i r, applied to the neighboring pixels of the current block, how to perform the filtering process, and the like.
20 EOlbO]
The filter deci -31 on uni t 1 '^ 3 si.ipp] .i es the control infornijJl ion w}i i r:h I'orM r<>l ^ exe£:vition of the f.i].tering prorr.r.r. I o l.he t nw p^.-i.^ Ti 1 t.er iini »- ] 4 4 and the bi-linear f-i 1 Ler un i I. 14^ -^cforfi i riy U) Uie deci s i on . Namely, the
25 filter i:3e<"i.'il(>ri un i L I'J3 Hiippl ieM i'ontrol i nf ormatj.on i nd i <:rfLiny hotJ to perform l.he f ill.er i ny pc oeess to the pr(n:erising unit selected beLijeetn Llie lou pm-a-: fi 1 l:er uni t 1^4 and the bi-linear filter unit l^h ancJ siippl i e.'j control information indicating to stop the filLt^ring
30 process (that is, control information indicating not. to perform the filter.ing process) to the non-selected

5..^'
^P3Jjl03?WO00
proce^]sing unit. [01511
The low pass filter urn l: 144 pe^rforni.s the low pass riiterincf proccGS dciscribt^d wi Lh refEjrenee to Fig. *j on b llie neighborinq pixels nf l.he ciirienl: block supplied troiti the nei (jbhoring pixel deteriviining unit I 'i?. accordi ng tn tht-' control information supp.l .i ed from (.hE^ f LlL.F^r decision uni I. 1^!S. The low pass filter uni t 14^ supfil i-f..^ I.he neighborinq pixel ^ (afl cr- EiH.ering-process neighboring
10 pixels) which ore appl i ed vjiS.h the low pass filtering process to I he iiilEa prediction unit 113. [015?]
'I'he hi-llnear filter unit I 45 performs t.he hi--.1 inear filtering pr.occ^ij described with reference to
IS Formulas (5) to (9) on L.lie neighboring pixels of the current b.l oek siipp.l ied from the neighboring pixel deterraininq uni\ 1A2 according to the control information suppl i f!d Ecoiii the filter decision unit 143. "I'be bi IT nctir filter unit 145 supplies the nei qhbori jiq pi xel y,
20 (after- rilLering-process neighboring p.i xel s) vjhi eh ^re
,-ipplicd with the bi-1 inear ri 1 L.ei-i nc; pro<:es,f, to the intra
prediction un.i t 113.
[Glb3]
By domq so, Ube Lhreslioid ualue setting unit 121
2b can set the threshold Vrtlue according to the bit depth of the image dni.^ [appropriate to the bit depth); the filter! rig processing unit 122 can select the fi.1 ter iippropriai.e to the bit depth of the image data and perform the filtering process on the neighbori ng pi xels
30 by using the filter j and the .i ntrrt. predi eUi on uni l. 11.3
r:^[i generate the predicted imaqe by usi mj I.he neighboring

pixels appl "i ed ivi l.h the f iltcrinq prci^cc-mi appropriate to the bit depL.h of the imaqe data- Nrimeiy, by doing 30, the irn^ge encoding dcvi ce lOH crfti .■suppress a deLerioration in "inatji-: (siiiaiil.y of the decoded imrtije. 5 [01541