融合面部信息的手語識別方法研究.pdf

1、大連海事大學碩士學位論文融合面部信息的手語識別方法研究姓名：梁芳申請學位級別：碩士專業(yè)：計算機科學與技術指導教師：王春立20090601英文摘要AbstractSignlanguage，asakindofmoststructuredgesture，isregardedasanindispensablemeansofeverydaycommunicationforthedea￡Asoneofthemostimportantpartsofh

2、uman—computerinteractiontheresearchandimplementationofsignlanguagerecognitionhasimportantacademicvalueaswellasbroadapplicationaspectAsthesignlanguagerecognitionisbasedonvisualsystem，itCanprovidemorenaturalandconvenienthu

3、mancomputerinteractionfunction，SOitisraisingmoreandmoreattentionforresearchersandhasbeenimprovedalotAfteranalysisthecurrentresearchStatuS，itisfoundthattheresearchesmainlyfocusonthehandinformationprocessing，modulecreating

4、andrecognition，butseldomonthehandlanguageisabodylanguagebasedonhandstyle，armmovement，supportedbythefacialexpression，lipsactingandotherbodyposturetodeliverconceptionThehandgestureandfacialexpressionplayakeypointatsignlang

5、uagerecognitionAndtheexperimentalresultsshowthatwhenthereisonlygesturewithoutexpressions，thecontentcanbeunderstoodbypeoplenomorethan60％So，basedonthehandsignlanguagerecognitiononlyusinghandcharacteristics，thispapercombine

6、sthecharacteristicsoffacialexpressionsinsignlanguagerecognition，viewtoimprovingtheefficiencyofsignlanguagerecognition。Indetail，themainresearchCanbedescribedasfollows：1ThisarticlechoosethecolorspaceYCbCrtosegmenttheimagew

7、iththebrightnessandchromaticityinformation，replacingoriginalRGBspacemethod，andreceivethemorerobustsegmentationeffectMeanwhile，thispaperextractthehandsregionbyUSeofskincolormodelandcombiningregionalconnection，andthenextra

8、ctthecharacteristicsofhandswiththemethodofellipsefitting2ProposedandimplementedthatfacialexpressionrecognitionintegrateswiththesignlanguageThisarticlefirstnormalizepositionedeyeandmouthareaintermsofsizeandgray—scale，sele