<?xml version="1.0" encoding="UTF-8"?><?xml-stylesheet type="text/xsl" href="static/style.xsl"?><OAI-PMH xmlns="http://www.openarchives.org/OAI/2.0/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/ http://www.openarchives.org/OAI/2.0/OAI-PMH.xsd"><responseDate>2026-09-18T17:38:48Z</responseDate><request verb="GetRecord" identifier="oai:publikationen.uni-tuebingen.de:10900/167383" metadataPrefix="mets">https://publikationen.uni-tuebingen.de/oai/request</request><GetRecord><record><header><identifier>oai:publikationen.uni-tuebingen.de:10900/167383</identifier><datestamp>2025-07-01T01:01:09Z</datestamp><setSpec>com_10900_42126</setSpec><setSpec>col_10900_42133</setSpec></header><metadata><mets xmlns="http://www.loc.gov/METS/" xmlns:doc="http://www.lyncode.com/xoai" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xlink="http://www.w3.org/1999/xlink" xsi:schemaLocation="http://www.loc.gov/METS/ http://www.loc.gov/standards/mets/mets.xsd" PROFILE="DSpace METS SIP Profile 1.0" TYPE="DSpace ITEM" ID="&#xa;&#x9;&#x9;&#x9;&#x9;DSpace_ITEM_10900-167383" OBJID="&#xa;&#x9;&#x9;&#x9;&#x9;hdl:10900/167383">
<metsHdr CREATEDATE="2026-09-18T19:38:48Z">
<agent TYPE="ORGANIZATION" ROLE="CUSTODIAN">
<name>Publikationssystem UB Tuebingen</name>
</agent>
</metsHdr>
<dmdSec ID="DMD_10900_167383">
<mdWrap MDTYPE="MODS">
<xmlData xmlns:mods="http://www.loc.gov/mods/v3" xsi:schemaLocation="http://www.loc.gov/mods/v3 http://www.loc.gov/standards/mods/v3/mods-3-1.xsd">
<mods:mods xsi:schemaLocation="http://www.loc.gov/mods/v3 http://www.loc.gov/standards/mods/v3/mods-3-1.xsd">
<mods:name>
<mods:role>
<mods:roleTerm type="text">advisor</mods:roleTerm>
</mods:role>
<mods:namePart>Giese, Martin A. (Prof. Dr.)</mods:namePart>
</mods:name>
<mods:name>
<mods:role>
<mods:roleTerm type="text">author</mods:roleTerm>
</mods:role>
<mods:namePart>Stettler, Michael</mods:namePart>
</mods:name>
<mods:extension>
<mods:dateAccessioned encoding="iso8601">2025-06-30T11:59:22Z</mods:dateAccessioned>
</mods:extension>
<mods:extension>
<mods:dateAvailable encoding="iso8601"/>
</mods:extension>
<mods:originInfo>
<mods:dateIssued encoding="iso8601">2025-06-30</mods:dateIssued>
</mods:originInfo>
<mods:identifier type="uri">http://hdl.handle.net/10900/167383</mods:identifier>
<mods:abstract>Gesichtsausdr¨ ucke spielen eine entscheidende Rolle in der sozialen Interaktion zwi-&#xd;
schen Menschen im t¨ aglichen Leben. Die menschliche Wahrnehmung ist dabei außer-&#xd;
gew&#xd;
¨ ohnlich robust bei der Vearbeitung unbekannter Gesichter, und sogar gegen¨ uber nicht-&#xd;
menschlichen Kopfformen (z. B. Emojis, Cartoons). Der visuelle Mechanismus, der&#xd;
zur Wahrnehmung dynamischer Gesichtsausdr¨ ucke f¨ uhrt, ist weitgehend unklar. In die-&#xd;
ser Arbeit zielen wir darauf ab, zu verstehen, wie unser visuelles System dynamische&#xd;
Gesichtsausdr¨ ucke codiert. Unter Verwendung von Methoden aus der Computergrafik&#xd;
haben wir Stimuli entwickelt, um die menschliche Verhaltenswahrnehmung durch psy-&#xd;
chophysische Experimente zu untersuchen, und gleichzeitig elektrophysiologische Auf-&#xd;
zeichnungen von Rhesusaffen gesammelt. Wir nutzen diese Daten, um neue Modelle&#xd;
zu entwerfen, zu testen und zu validieren. In diesem Rahmen haben wir eine neuar-&#xd;
tige Architektur entwickelt, die auf der Normreferenzcodierung basiert. Diese wurde&#xd;
urspr&#xd;
¨ unglich zur Modellierung von Gesichtsidentit¨ at vorgeschlagen und beruht auf der&#xd;
Annahme dass Gesichter als richtungsspezifische Abweichung eines Referenzgesichts&#xd;
repr&#xd;
¨ asentiert sind. Zun¨ achst zeigen wir, dass die Normreferenzcodierung ein g¨ ultiger&#xd;
Mechanismus zur Codierung von Gesichtsausdr¨ ucken ist. Wir erweitern dann die Norm-&#xd;
referenzcodierung auf mehrere Referenzrahmen und erkennen, dass das Lernverfahren&#xd;
in diesem Fall in zwei Teile aufgeteilt werden kann. Diese Multi-Dom¨ anen- Normrefe-&#xd;
¨&#xd;
renzcodierung stellt einen dateneffizienten Mechanismus dar, der die&#xd;
Ubertragung von&#xd;
Gesichtsausdr¨ ucken ¨ uber grundlegend unterschiedliche Gesichtsformen mit nur einem&#xd;
Datenpunkt erm¨ oglicht. Wir erweitern unser Modell, um es auf gr¨ oßeren Datens¨ atzen zu&#xd;
testen, und zeigen, dass der Mechanismus auch dann robust ist und generalisiert. Schließ-&#xd;
lich stellen wir fest, dass die Normreferenzcodierung im Vergleich zu anderen Computer-&#xd;
¨&#xd;
Vision-Modellen eine gr¨ oßere&#xd;
Ahnlichkeit zur Wahrnehmung unserer menschlichen Pro-&#xd;
banden aufweist. Insgesamt beleuchtet unsere Arbeit einen in der Neurowissenschaft be-&#xd;
kannten Mechanismus,der jedoch bisher in der Computer Vision wenig erforscht wurde,&#xd;
und zeigt, wie er in diesem Kontext genutzt werden kann. Die Normreferenzcodierung&#xd;
hat das Potenzial, Computersysteme zu entwickeln, deren Wahrnehmung von Gesichts-&#xd;
ausdr¨ ucken der menschlichen Wahrnehmung ¨ ahnlicher ist als die aktueller Techniken.</mods:abstract>
<mods:abstract>Facial expressions play a crucial role in daily human social interactions. Humans exhibit&#xd;
a remarkably strong and innate ability to perceive both novel faces and non-human head&#xd;
shapes, such as emojis and cartoons. However, the visual mechanisms involved in per-&#xd;
ceiving dynamic facial expressions remain largely unclear. This thesis aims to unravel&#xd;
the processes by which our visual system encodes dynamic facial expressions.&#xd;
Employing computer graphics, we have crafted stimuli to delve into human behavioral&#xd;
perception through psychophysics experiments. Concurrently, we collect electrophysi-&#xd;
ological recordings from rhesus macaques. Leveraging these data sets, we design, test,&#xd;
and validate our models. Our innovative framework, based on norm-referenced encod-&#xd;
ing—a mechanism initially proposed for encoding facial identity—proves to be a valid&#xd;
mechanism for encoding facial expressions.&#xd;
Expanding norm-referenced encoding into multi-domain applications, we discover&#xd;
that utilizing an updatable reference frame allows us to split the learning procedure, mak-&#xd;
ing it a data-efficient mechanism. This enables the transfer of facial expressions across&#xd;
basic face shapes with only a single training data point. Extending our model to larger&#xd;
datasets, we demonstrate its robustness and generalization capabilities. Notably, norm-&#xd;
referenced encoding yields perceptual results closer to human perception compared to&#xd;
other computer vision models.&#xd;
In summary, our research sheds light on a mechanism well-known in neuroscience&#xd;
but relatively unexplored in computer vision. We demonstrate its potential application&#xd;
in developing computer systems that interact and understand human facial expressions&#xd;
more similarly to current techniques. Norm-referenced encoding holds promise for the&#xd;
advancement of computer systems in this domain.</mods:abstract>
<mods:language>
<mods:languageTerm authority="rfc3066">en</mods:languageTerm>
</mods:language>
<mods:accessCondition type="useAndReproduction">cc_by</mods:accessCondition>
<mods:titleInfo>
<mods:title>Investigation and Modelling of Dynamical Facial Expression Perception</mods:title>
</mods:titleInfo>
<mods:genre>PhDThesis</mods:genre>
</mods:mods>
</xmlData>
</mdWrap>
</dmdSec>
<amdSec ID="FO_10900_167383_1">
<techMD ID="TECH_O_10900_167383_1">
<mdWrap MDTYPE="PREMIS">
<xmlData xmlns:premis="http://www.loc.gov/standards/premis" xsi:schemaLocation="http://www.loc.gov/standards/premis http://www.loc.gov/standards/premis/PREMIS-v1-0.xsd">
<premis:premis>
<premis:object>
<premis:objectIdentifier>
<premis:objectIdentifierType>URL</premis:objectIdentifierType>
<premis:objectIdentifierValue>https://publikationen.uni-tuebingen.de/xmlui/bitstream/10900/167383/1/PhD_Thesis_accepted_title.pdf</premis:objectIdentifierValue>
</premis:objectIdentifier>
<premis:objectCategory>File</premis:objectCategory>
<premis:objectCharacteristics>
<premis:fixity>
<premis:messageDigestAlgorithm>MD5</premis:messageDigestAlgorithm>
<premis:messageDigest>7d63ea3048c3d984411e187c01d02ffa</premis:messageDigest>
</premis:fixity>
<premis:size>43042569</premis:size>
<premis:format>
<premis:formatDesignation>
<premis:formatName>application/pdf</premis:formatName>
</premis:formatDesignation>
</premis:format>
</premis:objectCharacteristics>
<premis:originalName>PhD_Thesis_accepted_title.pdf</premis:originalName>
</premis:object>
</premis:premis>
</xmlData>
</mdWrap>
</techMD>
</amdSec>
<amdSec ID="FT_10900_167383_3">
<techMD ID="TECH_T_10900_167383_3">
<mdWrap MDTYPE="PREMIS">
<xmlData xmlns:premis="http://www.loc.gov/standards/premis" xsi:schemaLocation="http://www.loc.gov/standards/premis http://www.loc.gov/standards/premis/PREMIS-v1-0.xsd">
<premis:premis>
<premis:object>
<premis:objectIdentifier>
<premis:objectIdentifierType>URL</premis:objectIdentifierType>
<premis:objectIdentifierValue>https://publikationen.uni-tuebingen.de/xmlui/bitstream/10900/167383/3/PhD_Thesis_accepted_title.pdf.txt</premis:objectIdentifierValue>
</premis:objectIdentifier>
<premis:objectCategory>File</premis:objectCategory>
<premis:objectCharacteristics>
<premis:fixity>
<premis:messageDigestAlgorithm>MD5</premis:messageDigestAlgorithm>
<premis:messageDigest>33c802f987330f5bd26c962eb1e150be</premis:messageDigest>
</premis:fixity>
<premis:size>249169</premis:size>
<premis:format>
<premis:formatDesignation>
<premis:formatName>text/plain</premis:formatName>
</premis:formatDesignation>
</premis:format>
</premis:objectCharacteristics>
<premis:originalName>PhD_Thesis_accepted_title.pdf.txt</premis:originalName>
</premis:object>
</premis:premis>
</xmlData>
</mdWrap>
</techMD>
</amdSec>
<fileSec>
<fileGrp USE="ORIGINAL">
<file ID="BITSTREAM_ORIGINAL_10900_167383_1" MIMETYPE="application/pdf" SEQ="1" SIZE="43042569" CHECKSUM="7d63ea3048c3d984411e187c01d02ffa" CHECKSUMTYPE="MD5" ADMID="FO_10900_167383_1" GROUPID="GROUP_BITSTREAM_10900_167383_1">
<FLocat xlink:type="simple" LOCTYPE="URL" xlink:href="https://publikationen.uni-tuebingen.de/xmlui/bitstream/10900/167383/1/PhD_Thesis_accepted_title.pdf"/>
</file>
</fileGrp>
<fileGrp USE="TEXT">
<file ID="BITSTREAM_TEXT_10900_167383_3" MIMETYPE="text/plain" SEQ="3" SIZE="249169" CHECKSUM="33c802f987330f5bd26c962eb1e150be" CHECKSUMTYPE="MD5" ADMID="FT_10900_167383_3" GROUPID="GROUP_BITSTREAM_10900_167383_3">
<FLocat xlink:type="simple" LOCTYPE="URL" xlink:href="https://publikationen.uni-tuebingen.de/xmlui/bitstream/10900/167383/3/PhD_Thesis_accepted_title.pdf.txt"/>
</file>
</fileGrp>
</fileSec>
<structMap TYPE="LOGICAL" LABEL="DSpace Object">
<div TYPE="DSpace Object Contents" ADMID="DMD_10900_167383">
<div TYPE="DSpace BITSTREAM">
<fptr FILEID="BITSTREAM_ORIGINAL_10900_167383_1"/>
</div>
</div>
</structMap>
</mets></metadata></record></GetRecord></OAI-PMH>