TheAndroidPlotViewlibrarycanbeusedtoplotsignalsonAndroiddevices.ItisimplementedasanAndroidcontrolandhasnoexternaldependencies.
ThiscodeaccompaniesthepaperCameraCalibration:ActiveversusPassiveTargetsbyChristophSchmalz,FrankForsterandElliAngelopoulou,(OpticalEngineering,November2011)
Activetargetscanbedigitaldisplaysore-paper.Theyhaveseveraladvantages.Digitaldisplaysarenearlyubiquitous,sodedicatedcalibrationtargetswithafixedcheckerboardordotarraypatternbecomeunnecessary.Handlingiseasyasmostdisplayshaveatilt/swivelbasetosetupdifferenttargetposes.Also,thetargetdoesnothavetobeinfocustoobtainexcellentcalibrationresults,whichisespeciallyusefulforcalibratingcameraswithalargefieldofview,sincethetargetcanbemovedclosertothelens.Mostimportantly,activetargetsenableahigherlocalizationaccuracyforthecalibrationmarksthantraditionalpassivetargets.Thistranslatesintoahigheraccuracyinapplicationslikeforexample3Dreconstructiontasks.
Seethepaperandtheincludedreadmeforfurtherdetails.
Theareaofapplicationisthe3-Dand4-Dreconstructionofcontrastedcardiacvessels,e.g.thecoronaryarteriesusingC-armCT(rotationalangiography).Variousmethodsexistinliteraturewithlotsofniceresults.However,theresultscanvarysignificantlybetweenphantomandrealclinicaldata.
Therefore,weprovide:
ThisisaC++libraryforgeneral-purpose,unconstrainednon-linearoptimizationontheGPUusingNVIDA'sCUDA.ItimplementstheL-BFGS(“Limited-memoryBroyden-Fletcher-Goldfarb-Shanno“)method,apopularQuasi-Newtonoptimizationmethodwithalowmemoryfootprint.
ThisisaC++implementationoftheframeworkproposedinourpaper"GPUAcceleratedTime-of-FlightSuper-ResolutionforImage-GuidedSurgery".Itemploysmaximumaposteriori(MAP)estimationtoobtainanimprovedimageofhigherresolutionfrommultiplelow-resolutionimages.Ourimplementationacceleratessuper-resolutiononthegraphicsprocessingunit(GPU)usingNVIDIA'sCUDAplatform.
EpipolarConsistencyhasbeenshowntobeapowerfultoolincalibrationandmotioncorrectionforflatpaneldetectorCT(FD-CT)andotherX-raybasedapplications.Theepipolarconsistencyconditions(ECC)applytoprojectiondatadirectlyandcanbeusedtocorrectfor3Dparameterswithout3Dreconstruction.Thecomputationofanepipolarconsistencymetricisreal-timecapableforsmallsetsofimagesandcanbeparallelizedforlargesetsofimages.Thisisanopen-sourceimplementationinOctave/MATLAB,whichcaneasilybeportedtoyourfavoritesoftwareenvironment.
Embeddedmicrocontrollersareemployedinanincreasingnumberofapplicationsasatargetfortheimplementationofclassificationsystems.Thisistrue,forexample,forthefieldsofsports,automotive,andmedicalengineering.However,importantchallengesarisewhenimplementingclassificationsystemsonembeddedmicrocontrollers,whichismainlyduetolimitedhardwareresources.
DynaSLisacollectionoftoolstoanalyzesamplingstrategiesandreconstructionalgorithmsforT1-weightedDCE-MRI.Itcanbeusedto:
Gerbilisanopen-sourceprojectintendedforresearchersworkingwithmultispectralorhyperspectralimages,orresearchingandeducatingincolorandreflectance.Thesoftwareconsistsofanewinteractivevisualizationtooltargetedatmultispectralandhyperspectralimagedata,andatoolboxofcommonalgorithms,e.g.forsegmentation.
HeartyisanapplicationforAndroid-basedmobiledevicesthatallowsreal-timeelectrocardiogram(ECG)monitoringandautomatedarrhythmiadetectionbyanalyzingECGparameters.
Furthermore,datasetsfortheevaluationofdenoisingalgorithms,whichweusedinourpapers,maybedownloaded.
ArrhythmiadetectionalgorithmsrequiretheexactandinstantaneousdetectionoffiducialpointsintheECGsignal.Thesefiducialpoints(QRS-complex,P-andT-waves)correspondtodistinctcardiaccontractionphases.
WeimplementedinMatlabthreealgorithms:oneisbasedonalinefittingapproach,anexpeditiousversionofaWavelet-basedECGdelineationalgorithm,andafastnaivefiducialpointsdetectionalgorithm.
InthisMATLABtoolbox,weprovidetheimplementationsofseveralstate-of-the-artalgorithmsaswellasnovelmethodsdevelopedinourprojectsonmuti-framesuper-resolution.Thealgorithmscovergeneral-purposereconstructionalgorithmsthatexploitsubpixelmotiontogainsuper-resolveddataandtailor-madesolutionsforspecificapplications(withfocusonapplicationsinmedicalimaging).
OCTSEG(OpticalCoherenceTomographySegmentationandEvaluationGUI)isagraphicaluserinterfacewritteninMATLABforresearchpurposes.
RabbitCTisanopen-sourcebenchmarkforC-armCTFBPreconstruction.
ORCA-SPOTisadeeplearningtoolkitimplementedinPyTorchanddesignedtosegmentinterestinganimalsounds(killerwhalesignals)outoflarge,noise-heavy,andheterogenousbioacousticaudiocorpora.ORCA-SPOTcanbeadaptedtootheranimalspecies.ThecodewillbepubliclyavailableuponrequestonlyattheendofOctober2019.ForquestionspleasecontactChristianBergler(christian.bergler@fau.de).