简介：

简介：

简介：在计算机化的断层摄影术(CT)的膀胱的分割想象是在放射治疗计划前列腺癌症的重要的步。我们在场自动地描出的一个新分割计划在与三个学生一起的CT图象的膀胱轮廓走。首先，我们使用吝啬的移动算法获得包含膀胱的不平的轮廓的一幅聚类的图象，它然后被使用一个成长区域的算法，起始的种子点从扫描进程的一个线每篇文字题目下作者的署名选择了在第二步提取。第三步是用转动球算法更精确地精制膀胱轮廓。这些步然后被扩大以一种slice-by-slice方式分割膀胱体积。获得的结果与由放射肿瘤学家的用手的分割相比。敏感，特性，积极预计用价值，否定预计用价值，和Hausdorff距离的平均价值分别地是86.5%，96.3%，90.5%，96.5%，和2.8象素。结果证明膀胱能精确地被分割。

简介：Severalfeaturesofretinalvesselscanbeusedtomonitortheprogressionofdiseases.Changesinvascularstructures,forexample,vesselcaliber,branchingangle,andtortuosity,areportentsofmanydiseasessuchasdiabeticretinopathyandarterialhypertension.Thispaperproposesanautomaticretinalvesselsegmentationmethodbasedonmorphologicalclosingandmulti-scalelinedetection.First,anilluminationcorrectionisperformedonthegreenbandretinalimage.Next,themorphologicalclosingandsubtractionprocessingareappliedtoobtainthecruderetinalvesselimage.Then,themulti-scalelinedetectionisusedtofinethevesselimage.Finally,thebinaryvasculatureisextractedbytheOtsualgorithm.Inthispaper,forimprovingthedrawbacksofmulti-scalelinedetection,onlythelinedetectorsat4scalesareused.Theexperimentalresultsshowthattheaccuracyis0.939forDRIVE(digitalretinalimagesforvesselextraction)retinaldatabase,whichismuchbetterthanothermethods.

简介：ThereisdifficultyfordistinguishingofriverandshadowinSyntheticApertureRadar(SAR)images.AmethodofriversegmentationinSARimagesbasedonwaveletenergyandgradientisproposedinthispaper.Itmainlyincludestwoalgorithms:coarsesegmentationandrefinedsegmen-tation.Firstly,Theriverregionsarecoarselysegmentedbythewaveletenergyfeature,andthenrefinedsegmentedaccuratelybythegradientthresholdwhichisgotadaptively.Theexperimentalresultsshowthevalidityofthemethod,whichprovidesagoodfoundationfortargetsdetectionabovetheriver.

简介：Inordertounderstandthedevelopmentofstemcellsintospecializedmaturecellsitisnecessarytostudythegrowthofcellsinculture.Forthispurposeitisveryusefultohaveanefficientcomputerizedcelltrackingsystem.Inthispaperaprototypesystemfortrackingneuralstemcellsinasequenceofimagesisdescribed.Inordertogetreliabletrackingresultsitisimportanttohavegoodandrobustsegmentationofthecells.Toachievethiswehaveimplementedthreelevelsofsegmentation.Theprimarylevel,appliedtoallframes,isbasedonfuzzythresholdandwatershedsegmentationofafuzzygrayweighteddistancetransformedimage.Thesecondlevel,appliedtodifficultframeswherethefirstalgorithmseemstohavefailed,isbasedonafastgeometricactivecontourmodelbasedonthelevelsetalgorithm.Finally,theautomaticsegmentationresultonthecrucialfirstframecanbeinteractivelyinspectedandcorrected.Visualinspectionandcorrectioncanalsobeappliedtootherframesbutthisisgenerallynotneeded.Forthetrackingallcellsareclassifiedintoinactive,active,dividingandclusteredcells.Differentalgorithmsareusedtodealwiththedifferentcellcategories.Aspecialbacktrackingstepisusedtoautomaticallycorrectforsomecommonerrorsthatappearintheinitialforwardtrackingprocess.

简介：语义图象分割是一项任务为每个图象象素预言一个范畴标签。它的关键挑战是设计一个强壮的特征代表。在这份报纸，我们作为特征表示熔化神经网络(CNN)展示的层次convolutional和基于区域的特征。层次特征包含更全球的信息，当基于区域的特征包含更多的本地信息时。这些二种特征的联合显著地提高特征表示。然后，熔化特征被用来训练一个softmax分类器生产每象素标签任务概率。并且一块充分连接的有条件的随机的地(CRF)被用作一个processing以后方法改进标记的一致性。我们进行实验在上筛流动数据集。象素精确性和班精确性分别地是84.4%和34.86%。

简介：质地图像分割基于非线性的散开被介绍。质地的规模能在非线性的散开的过程期间被测量。有保存的边的一幅光滑的5隧道向量图象，紧张镇静，可伸缩，质地的取向想象，能被联合非线性的散开完成。一个多信道的统计区域活跃轮廓被采用分割这幅向量图象。因为参数不对不同质地图象敏感，方法能被看作一种无指导的分割。试验性的结果在质地图象的半自动的抽取显示出它的高效率。

简介：Imagesegmentation,asabasicbuildingblockformanyhigh-levelimageanalysisproblems,hasattractedmanyresearchattentionsoveryears.Existingapproaches,however,aremainlyfocusingontheclusteringanalysisinthesinglechannelinformation,i.e.,eitherincolororspatialspace,whichmayleadtounsatisfactorysegmentationperformance.Consideringthespatialandcolorspacesjointly,thispaperproposesanewhierarchicalimagesegmentationalgorithm,whichalternatelyclusterstheimageregionsincolorandspatialspacesinafinetocoarsemanner.Withoutlosingtheperceptualconsistence,theproposedalgorithmachievesthesegmentationresultusingonlyveryfewnumberofcolorsaccordingtouserspecification.

简介：Thispaperpresentsanovelmethodofspotaddressingandsegmentationabouttheforegroundsegmentationofmicroarrayimage.Inthispaper,aspotaddressingmethodbasedonparticleswarmoptimization(PSO),algorithmisproposedtohaveafurthersearchforthecentercoordinateandradiusofthespotwhoseregionisdeterminedbytheprojectionmethod.Then,aforegroundsegmentationmethodisputforwardtomakethespotforegroundsegmentationbasedonthecentercoordinateandradiusofthespot.Thespotaddressingandsegmentationexperimentsonsyntheticandrealmicroarrayimagesshowthattheproposedmethodiseffectiveandfeasiblefortheforegroundsegmentationofmicroarrayimage.

简介：

简介：Thispaperconiesupwithanewideathattriestooptimizetheperformanceofimagesegmentationalgorithmsbasedontheirobjectiveevaluationknowledge.Aprototypeexpertsystemisdesignedandimplementedaccordingtothisidea.Kxperimentalresultsindicatethatthissystemisfeasibleandpractical.Thisideaalsooffersanefficientapproachforalgorithmoptimizationofimageprocessingandanalysis.

简介：这份报纸试图介绍一种交互颜色自然图象分割方法。这个方法由使用非线性的紧缩的结构张肌(NCST)然后使用GrabCut方法获得分割提取图象的特征。这个方法不仅认识到质地信息和颜色信息的非参量的熔化，而且改进计算的效率。然后，改进GrabCut算法被用来评估前景目标分割。以便计算简洁和效率，这份报纸也扩大构造的Gaussian混合模型(GMM)到张肌空间的GrabCut上的底，和使用Kullback-Leibler(KL)分叉而不是平常的Riemannian几何学。最后，一个重复集中标准被建议与满足的分割精确性戏剧性地减少GrabCut算法的重复的时间。在在合成质地图象和自然图象上进行很多实验以后，结果证明这个方法有更精确的分割效果。

简介：

简介：

简介：Osteosarcomaisprimarymalignantneoplasmsderivedfromcellsofmesenchymalorigin,andoftenhasdistinctphenotypesatdifferentstages.Thelocationoftumorandreactionzonecanbeidentifiedbyanexpertinmagneticresonanceimaging(MRI),withMRIbeingoneofthechoicesforevaluatingtheextentofosteosarcoma.However,itisstillachallengetoautomaticallyextracttumorfromitssurroundingtissuesbecauseoftheirlowintensitydifferencesinMRI.WeinvestigatedanapproachbasedonZernikemomentandsupportvectormachine(SVM)forosteosarcomasegmentationinT1-weightedimage(TIWI).Firstly,thedifferentordermomentsaroundeachpixelarecalculatedinsmallwindows.Secondly,thegrayscaleandthemodulevaluesofdifferentordermomentsareusedasatexturefeaturevectorwhichisthenusedasthetrainingsetforSVM.Finally,anSVMclassifieristrainedbasedonthissetoffeaturestoidentifytheosteosarcoma,andthesegmentedtumortissueisrenderedin3Dbytheraycastingalgorithmbasedongraphicsprocessingunit(GPU).TheperformanceofthemethodisvalidatedonT1WI,showingthatthesegmentationmethodhasahighsimilarityindexwiththeexpert’smanualsegmentation.

简介：Thequantumtheoryapplicationisahotresearchareainrecentyears,especiallythetheoryofquantummechanics.Inthispaper,wefocusontheresearchofimagesegmentationbasedonquantummechanics.Firstly,thetheoryofquantummechanicsisintroduced;afterwards,areviewofimagesegmentationmethodsbasedonquantummechanicsispresented;andfinally,thecharacteristicsaboutthequantummechanicsappliedtoimageprocessingareconcluded.Twomainresearchtopicsarediscussedinthispaper.Oneistoemphasizethatquantummechanicscanbeappliedindifferentresearchareas,suchasimagesegmentation,andthesecondistoconcludesomemethodsinimagesegmentationandgivesomesuggestionsforpossiblenovelmethodsbyapplyingquantummechanicstheory.Asasummary,thisisareviewpaperwhichpresentssomemethodsbasedonthefeasibletheoryinquantummechanicsaimingatachievingabetterperformanceinimagesegmentation.

简介：Inthispaper,theX-raynondestructivetestmethodofsmalldefectsinprecisionweldmentswithcomplexstructurewaspresented.Toresolvethedifficultyofdefectsegmentationinvariablegreyimage,theimageprocessingbasedonVisualBasicprogrammingmethodwasadopted.ThemethodsofautomaticcontrastandpartialgreystretchwereusedtoenhancetheX-raydetectionimagewhichhasrelativelylowcontrast,thenautomaticthresholdmethodwascarriedouttosegmentthetwohighintensityzones,andweldzoneswhichcontainthesmalldefectswasextracted.Smoothingandsharpenprocessingwereproceededontheextractedweldzones,andsmalldefectsinX-raydetectionimageofweldmentswithcomplexstructureweresegmentedbyusingthemethodofbackgroundsubtractionintheend.Theeffectsofrasterwereeliminated,andbecauseofthattheimageprocessingwasonlyproceededontheextractedweldzones,thecalculatedspeedusingtheaboveprovidedalgorithmwasimproved.

简介：

简介：