资源描述：
《A Comnpressive Sensing Image Compression Algorithm Using Quantized DCT and Noiselet Information》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。
1、ACOMPRESSIVESENSINGIMAGECOMPRESSIONALGORITHMUSINGQUANTIZEDDCTANDNOISELETINFORMATIONJiangtaoWen1,ZhuoyuanChen1,YuxingHan2,JohnD.Villasenor2andShiqiangYang1TsinghuaUniversity1,UniversityofCalifornia,LosAngeles2ABSTRACTthatissharedbybothencoderanddecodercanbeexploitedtolowertheoverallbitrate.Whilethisa
2、pproachisused,forInspiredbyrecenttheoreticaladvancesincompressivesens-example,inthespatialandtemporalpredictionsinH.264,iting(CS),weproposeanewframeworkthatcombinesthehasnotpreviouslybeenappliedtoCS.Second,weconsiderclassicallocaldiscretecosinetransformusedinimagecom-quantization,whichisinherentinan
3、ydigitalrepresentationpressionalgorithmssuchasJPEGwithaglobalnoiseletmea-ofanimage.Third,weconsiderbitrate,whichisofcourse,surewhichissolvedusingsecondorderconeprogrammingalongwithquality,oneofthemostimportantdeterminantsof(SOCP).compressionalgorithmperformance.IndexTerms—compressivesensing,SOCP,ima
4、gecom-pression,exp-Golombcoding,noiselet,quantization2.MATHEMATICALFRAMEWORKFORCOMPRESSIVESENSINGOFIMAGES1.INTRODUCTIONAsexplainedin[7][1][2][8]andelsewhere,givenasig-Compressivesensing(CS)providesaformalizedmathemat-nalxofdimensionn,CSaimstoprovideareconstructionicalframeworkforexploitingtheinheren
5、tlysparsenatureofbasedonmeasurementsyobtainedusingasensingbasisΦ=commonlyencounteredsignalsandhasbeenthesubjectof[φ...φ].Theestimateofx,denotedx∗isobtainedthrough1nsignificantactivityinrecentyears.Despitethesignificantx∗=Ψz∗,whereΨ=[ψ...ψ]isarepresentationba-1namountofattentionthathasbeengiventotheore
6、ticalas-sis(whichisgenerallydifferentfromthesensingbasisΦ=pectsofCS,practicalimagecompressionisstilldominated[φ...φ])andz∗isthesolutionofaconvexoptimization1nbyJPEGandJPEG-2000.Thesestandards,whileexploitingproblemasdescribedbelow.GivenknowledgeofΦandΨ,someofthesameunderlyingopportunitiesforeliminat
7、ingtheencodingprocessconsistsofprojectingxontoΦusingredundancyasCS,donotspecificallytakeadvantageoftheaninnerproduct,anddecodingrequiressolvingz∗,typicallytheoreticalfoundationsprovidedbyCS.throughSOCP
