简介:Facede-identificationhasbecomeincreasinglyimportantastheimagesourcesareexplosivelygrowingandeasilyaccessible.Theadvanceofnewfacerecognitiontechniquesalsoarisespeople'sconcernregardingtheprivacyleakage.Themainstreampipelinesoffacede-identificationaremostlybasedonthek-sameframework,whichbearscritiquesofloweffectivenessandpoorvisualquality.Inthispaper,weproposeanewframeworkcalledPrivacy-Protective-GAN(PP-GAN)thatadaptsGAN(generativeadversarialnetwork)withnovelverificatorandregulatormodulesspeciallydesignedforthefacede-identificationproblemtoensuregeneratingde-identifiedoutputwithretainedstructuresimilarityaccordingtoasingleinput.Weevaluatetheproposedapproachintermsofprivacyprotection,utilitypreservation,andstructuresimilarity.Ourapproachnotonlyoutperformsexistingfacede-identificationtechniquesbutalsoprovidesapracticalframeworkofadaptingGANwithpriorsofdomainknowledge.
简介:Denovosequencingisoneofthemostpromisingproteomicstechniquesforidentificationofproteinposttranslationmodifications(PTMs)instudyingproteinregulationsandfunctions.WehavedevelopedacomputertoolPRIMEforidentificationofbandyionsintandemmassspectra,akeychallengingproblemindenovosequencing.PRIMEutilizesafeaturethationsofthesameanddifferenttypesfollowdifferentmass-differencedistributionstoseparatebfromyionscorrectly.Wehaveformulatedtheproblemasagraphpartitionproblem.Alinearinteger-programmingalgorithmhasbeenimplementedtosolvethegraphpartitionproblemrigorouslyandefficiently.TheperformanceofPRIMEhasbeendemonstratedonalargeamountofsimulatedtandemmassspectraderivedfromYeastgenomeanditspowerofdetectingPTMshasbeentestedon216simulatedphosphopeptides.