染色體識別算法的研究.pdf

1、Y1041833申淆同濟大學理學碩士學位論文染色體識別算法的研究培養(yǎng)單位化學系一級學科：化學=級學科：分析化學研究生：郭翔指導教師叢培盛副教授AbstractABSTRACTIntheclinicalresearchofcellgenetics，c蛔omosomeanalysisisoneofthemostessentialtasksofgeneticsyndromediagnosisCorrectkaryotypeimage，whic

2、hcouldbeobtainedthroughchromosomeimageprocessing，togetherwithprecisec柵omosomeanalysisandrecognitionpmvidesvaluableinformationfordoctorinthediagnosingofhumanillnessThispapermainlystudiesimagepreprocessing，featureextractio

3、nandclassificationanalysis，accordinglyrealizesthechromosomerecognitionBeforeimagepreprocessing，humansegmentationneedtobeappliedtothewholechromosomeimageinordertogetsin西echromosomes，andthen，straightenupandnormalizethechro

4、mosomesbyimageprocessingsoftwareDuringthepreprocessing，clearing“saltandpepper’noisebymedianfiltering，binarizingimage，smoothingboundarybymorpholo西calcleaningandgraystretchingaltogetherleadtoaperfectimageThelengthofchromos

5、ome，centromericindexandGbandservingasthreepatternfeaturesareusedforchomosomerecognitionThelengthandcentromericindexcanbemeasuredbyimageprocessingsoRware，andG—bandstandsfortheaveragegrayresultingfromthemediaaxisInthestudy

6、ofc柵omosomerecognitionalgorithm，weuseaveragegraydistributionprofiletodenoteGbandbringforwardaconceptofchromosomefingerprint，selectcosineofanglemethodtocalculatethesimilarityandmatchthehomologouschoromosomesMeanwhile，bett

7、erresultsrootintheresolveofdifferentlengthsofchromosomesbyprofilestraightening，aswellasthesolvingofpe酞shiftbypmfilemovingIllustratingGbandbygraygradientprofile，ratherthanaveragegraydistributionprofile，ismuchmoreeffective