j. j. rissanen generalized kraft inequality and arithmetic coding

j. j. rissanen generalized kraft inequality and arithmetic coding

ID:34424179

大小:400.68 KB

页数:6页

时间:2019-03-06

j. j. rissanen generalized kraft inequality and arithmetic coding_第1页
j. j. rissanen generalized kraft inequality and arithmetic coding_第2页
j. j. rissanen generalized kraft inequality and arithmetic coding_第3页
j. j. rissanen generalized kraft inequality and arithmetic coding_第4页
j. j. rissanen generalized kraft inequality and arithmetic coding_第5页
资源描述:

《j. j. rissanen generalized kraft inequality and arithmetic coding》由会员上传分享,免费在线阅读,更多相关内容在教育资源-天天文库

1、J.J.RissanenGeneralizedKraftInequalityandArithmeticCodingAbstract:Algorithmsforencodinganddecodingfinitestringsoverafinitealphabetaredescribed.ThecodingoperationsarearithmeticinvolvingrationalnumbersliasparameterssuchthatZi2"i52".Thiscodingtechniquereq

2、uiresnoblocking,andtheper-symbollengthoftheencodedstringapproachestheassociatedentropywithinE.Thecodingspeediscomparabletothatofconventionalcodingmethods.IntroductionTheoptimalconventionalinstantaneousHuffmancodeThismeansthat,ifthestringsaregeneratedby

3、anin-[11foranindependentinformationsourcewithsymboldependentinformationsource,themeanofn"L(s)probabilities(p,,...,p,)maybeviewedasasolutiontoapproachestheentropyfunctionfromabovewithinantheintegerprogrammingproblem:Findmnaturalnum-errorE+O(1/n).berslia

4、slengthsofbinarycodewordssuchthatCipil,isThecodingoperationsarearithmetic,andtheyre-minimizedundertheconstrainingKraftinequalitysembletheconcatenationoperationsinconventionalcodinginthatthecodeofthestringsak,whereukisaI.newsymbol,isobtainedbyreplacingo

5、nlyafew(always1lessthansomefixednumber)oftheleft-mostbitsoftheThentheminimizedsumB,p,l,approximatestheShannon-coderepresentingsbyanewlongerstring.Asaresult,BoltzmannentropyfunctionH(p,,...,p,)=-zip,logthecodingoperationscanbeaccomplishedwithaspeedpifro

6、mabovewithanerrornomorethanone.Ifabettercomparabletothatofconventionalcoding.approximationisrequired,blockingisneeded;e.g.,akthTheprimaryadvantageoftheresulting"arithmeticextensionofthealphabetmustbeencoded,whichre-coding"isthatthereisnoblockingneedede

7、venfortheducestheleastupperboundoftheerrorto1/k[2].binaryalphabet.Inaddition,forsmallalphabetsizestheWedescribeanothercodingtechniqueinwhichmsizeoftablestobestoredandsearchedissmallerthanpositiverationalnumbersI,,...,1,areselectedsuchthecodewordtablesi

8、nconventionalcodingmethods.thatageneralizedKraftinequalityholds:ForabinaryalphabetaspecialchoiceofparametersI,andI,leadstoaparticularlysimplecodingtechnique2"i52-€,E>0.thatiscloselyrelatedtooneduetoFano[31.1Thecodingmethoddescribedherei

当前文档最多预览五页,下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页,下载文档查看全文
温馨提示:
1. 部分包含数学公式或PPT动画的文件,查看预览时可能会显示错乱或异常,文件下载后无此问题,请放心下载。
2. 本文档由用户上传,版权归属用户,天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容,确认文档内容符合您的需求后进行下载,若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误,付费完成后未能成功下载的用户请联系客服处理。