https://mw.hh.se/caisr/index.php?action=history&feed=atom&title=Publications%3ACross_Spectral_Periocular_Matching_using_ResNet_Features 2026-04-04T15:29:33Z Revision history for this page on the wiki MediaWiki 1.35.13 https://mw.hh.se/caisr/index.php?title=Publications:Cross_Spectral_Periocular_Matching_using_ResNet_Features&diff=4291&oldid=prev 2019-09-23T20:22:46Z

<p>Created page with &quot;&lt;div style=&#039;display: none&#039;&gt; == Do not edit this section == &lt;/div&gt; {{PublicationSetupTemplate|Author=Kevin Hernandez-Diaz, Fernando Alonso-Fernandez, Josef Bigun |PID=1348579 |...&quot;</p> <p><b>New page</b></p><div>&lt;div style=&#039;display: none&#039;&gt;<br /> == Do not edit this section ==<br /> &lt;/div&gt;<br /> {{PublicationSetupTemplate|Author=Kevin Hernandez-Diaz, Fernando Alonso-Fernandez, Josef Bigun<br /> |PID=1348579<br /> |Name=Hernandez-Diaz, Kevin (kevher) (Högskolan i Halmstad (2804), Akademin för informationsteknologi (16904), Halmstad Embedded and Intelligent Systems Research (EIS) (3938), CAISR Centrum för tillämpade intelligenta system (IS-lab) (13650));Alonso-Fernandez, Fernando (feralo) (0000-0002-1400-346X) (Högskolan i Halmstad (2804), Akademin för informationsteknologi (16904), Halmstad Embedded and Intelligent Systems Research (EIS) (3938), CAISR Centrum för tillämpade intelligenta system (IS-lab) (13650));Bigun, Josef (josef) (0000-0002-4929-1262) (Högskolan i Halmstad (2804), Akademin för informationsteknologi (16904), Halmstad Embedded and Intelligent Systems Research (EIS) (3938), CAISR Centrum för tillämpade intelligenta system (IS-lab) (13650))<br /> |Title=Cross Spectral Periocular Matching using ResNet Features<br /> |PublicationType=Conference Paper<br /> |ContentType=Refereegranskat<br /> |Language=eng<br /> |Journal=<br /> |JournalISSN=<br /> |Status=<br /> |Volume=<br /> |Issue=<br /> |HostPublication=<br /> |Conference=12th IAPR International Conference on Biometrics, Crete, Greece, June 4-7, 2019<br /> |StartPage=<br /> |EndPage=<br /> |Year=2019<br /> |Edition=<br /> |Pages=<br /> |City=<br /> |Publisher=<br /> |Series=<br /> |SeriesISSN=<br /> |ISBN=<br /> |Urls=<br /> |ISRN=<br /> |DOI=<br /> |ISI=<br /> |PMID=<br /> |ScopusId=<br /> |NBN=urn:nbn:se:hh:diva-40499<br /> |LocalId=<br /> |ArchiveNumber=<br /> |Keywords=<br /> |Categories=Signalbehandling (20205)<br /> |ResearchSubjects=<br /> |Projects=<br /> |Notes=<br /> |Abstract=&lt;p&gt;Periocular recognition has gained attention in the last years thanks to its high discrimination capabilities in less constraint scenarios than other ocular modalities. In this paper we propose a method for periocular verification under different light spectra using CNN features with the particularity that the network has not been trained for this purpose. We use a ResNet-101 pretrained model for the ImageNet Large Scale Visual Recognition Challenge to extract features from the IIITD Multispectral Periocular Database. At each layer the features are compared using χ &lt;sup&gt;2&lt;/sup&gt; distance and cosine similitude to carry on verification between images, achieving an improvement in the EER and accuracy at 1% FAR of up to 63.13% and 24.79% in comparison to previous works that employ the same database. In addition to this, we train a neural network to match the best CNN feature layer vector from each spectrum. With this procedure, we achieve improvements of up to 65% (EER) and 87% (accuracy at 1% FAR) in cross-spectral verification with respect to previous studies.&lt;/p&gt;<br /> |Opponents=<br /> |Supervisors=<br /> |Examiners=<br /> |Patent=<br /> |ThesisLevel=<br /> |Credits=<br /> |Programme=<br /> |Subject=<br /> |Uppsok=<br /> |DefencePlace=<br /> |DefenceLanguage=<br /> |DefenceDate=<br /> |CreatedDate=2019-09-04<br /> |PublicationDate=2019-09-04<br /> |LastUpdated=2019-09-23<br /> |diva=http://hh.diva-portal.org/smash/record.jsf?searchId=1&amp;pid=diva2:1348579}}<br /> &lt;div style=&#039;display: none&#039;&gt;<br /> <br /> == Keep all hand-made modifications below ==<br /> &lt;/div&gt;{{PublicationDisplayTemplate}}</div>

Slawek