Konx : cross-resolution image quality assessment

Lade...
Vorschaubild
Dateien
Wiedemann_2-1ujtqtmltl05g9.pdf
Wiedemann_2-1ujtqtmltl05g9.pdfGröße: 1.91 MBDownloads: 19
Datum
2023
Herausgeber:innen
Kontakt
ISSN der Zeitschrift
Electronic ISSN
ISBN
Bibliografische Daten
Verlag
Schriftenreihe
Auflagebezeichnung
ArXiv-ID
Internationale Patentnummer
Link zur Lizenz
Angaben zur Forschungsförderung
Deutsche Forschungsgemeinschaft (DFG): 251654672
Deutsche Forschungsgemeinschaft (DFG): 251654672
Projekt
Open Access-Veröffentlichung
Open Access Hybrid
Core Facility der Universität Konstanz
Gesperrt bis
Titel in einer weiteren Sprache
Publikationstyp
Zeitschriftenartikel
Publikationsstatus
Published
Erschienen in
Quality and User Experience. Springer. 2023, 8, 8. ISSN 2366-0139. eISSN 2366-0147. Available under: doi: 10.1007/s41233-023-00061-8
Zusammenfassung

Scale-invariance is an open problem in many computer vision subfields. For example, object labels should remain constant across scales, yet model predictions diverge in many cases. This problem gets harder for tasks where the ground-truth labels change with the presentation scale. In image quality assessment (IQA), down-sampling attenuates impairments, e.g., blurs or compression artifacts, which can positively affect the impression evoked in subjective studies. To accurately predict perceptual image quality, cross-resolution IQA methods must therefore account for resolution-dependent discrepancies induced by model inadequacies as well as for the perceptual label shifts in the ground truth. We present the first study of its kind that disentangles and examines the two issues separately via KonX , a novel, carefully crafted cross-resolution IQA database. This paper contributes the following: 1. Through KonX , we provide empirical evidence of label shifts caused by changes in the presentation resolution. 2. We show that objective IQA methods have a scale bias, which reduces their predictive performance. 3. We propose a multi-scale and multi-column deep neural network architecture that improves performance over previous state-of-the-art IQA models for this task. We thus both raise and address a novel research problem in image quality assessment.

Zusammenfassung in einer weiteren Sprache
Fachgebiet (DDC)
004 Informatik
Schlagwörter
Konferenz
Rezension
undefined / . - undefined, undefined
Forschungsvorhaben
Organisationseinheiten
Zeitschriftenheft
Datensätze
Zitieren
ISO 690WIEDEMANN, Oliver, Vlad HOSU, Shaolin SU, Dietmar SAUPE, 2023. Konx : cross-resolution image quality assessment. In: Quality and User Experience. Springer. 2023, 8, 8. ISSN 2366-0139. eISSN 2366-0147. Available under: doi: 10.1007/s41233-023-00061-8
BibTex
@article{Wiedemann2023-08-18cross-68952,
  year={2023},
  doi={10.1007/s41233-023-00061-8},
  title={Konx : cross-resolution image quality assessment},
  volume={8},
  issn={2366-0139},
  journal={Quality and User Experience},
  author={Wiedemann, Oliver and Hosu, Vlad and Su, Shaolin and Saupe, Dietmar},
  note={Article Number: 8}
}
RDF
<rdf:RDF
    xmlns:dcterms="http://purl.org/dc/terms/"
    xmlns:dc="http://purl.org/dc/elements/1.1/"
    xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
    xmlns:bibo="http://purl.org/ontology/bibo/"
    xmlns:dspace="http://digital-repositories.org/ontologies/dspace/0.1.0#"
    xmlns:foaf="http://xmlns.com/foaf/0.1/"
    xmlns:void="http://rdfs.org/ns/void#"
    xmlns:xsd="http://www.w3.org/2001/XMLSchema#" > 
  <rdf:Description rdf:about="https://kops.uni-konstanz.de/server/rdf/resource/123456789/68952">
    <void:sparqlEndpoint rdf:resource="http://localhost/fuseki/dspace/sparql"/>
    <dc:creator>Wiedemann, Oliver</dc:creator>
    <dc:contributor>Hosu, Vlad</dc:contributor>
    <dcterms:isPartOf rdf:resource="https://kops.uni-konstanz.de/server/rdf/resource/123456789/36"/>
    <dc:language>eng</dc:language>
    <dc:creator>Su, Shaolin</dc:creator>
    <dc:date rdf:datatype="http://www.w3.org/2001/XMLSchema#dateTime">2024-01-05T11:35:31Z</dc:date>
    <dcterms:title>Konx : cross-resolution image quality assessment</dcterms:title>
    <dcterms:available rdf:datatype="http://www.w3.org/2001/XMLSchema#dateTime">2024-01-05T11:35:31Z</dcterms:available>
    <dc:contributor>Su, Shaolin</dc:contributor>
    <dc:contributor>Saupe, Dietmar</dc:contributor>
    <dcterms:abstract>Scale-invariance is an open problem in many computer vision subfields. For example, object labels should remain constant across scales, yet model predictions diverge in many cases. This problem gets harder for tasks where the ground-truth labels change with the presentation scale. In image quality assessment (IQA), down-sampling attenuates impairments, e.g., blurs or compression artifacts, which can positively affect the impression evoked in subjective studies. To accurately predict perceptual image quality, cross-resolution IQA methods must therefore account for resolution-dependent discrepancies induced by model inadequacies as well as for the perceptual label shifts in the ground truth. We present the first study of its kind that disentangles and examines the two issues separately via KonX , a novel, carefully crafted cross-resolution IQA database. This paper contributes the following: 1. Through KonX , we provide empirical evidence of label shifts caused by changes in the presentation resolution. 2. We show that objective IQA methods have a scale bias, which reduces their predictive performance. 3. We propose a multi-scale and multi-column deep neural network architecture that improves performance over previous state-of-the-art IQA models for this task. We thus both raise and address a novel research problem in image quality assessment.</dcterms:abstract>
    <dcterms:hasPart rdf:resource="https://kops.uni-konstanz.de/bitstream/123456789/68952/1/Wiedemann_2-1ujtqtmltl05g9.pdf"/>
    <dc:contributor>Wiedemann, Oliver</dc:contributor>
    <foaf:homepage rdf:resource="http://localhost:8080/"/>
    <dc:creator>Hosu, Vlad</dc:creator>
    <dspace:isPartOfCollection rdf:resource="https://kops.uni-konstanz.de/server/rdf/resource/123456789/36"/>
    <dcterms:rights rdf:resource="http://creativecommons.org/licenses/by/4.0/"/>
    <bibo:uri rdf:resource="https://kops.uni-konstanz.de/handle/123456789/68952"/>
    <dc:creator>Saupe, Dietmar</dc:creator>
    <dcterms:issued>2023-08-18</dcterms:issued>
    <dspace:hasBitstream rdf:resource="https://kops.uni-konstanz.de/bitstream/123456789/68952/1/Wiedemann_2-1ujtqtmltl05g9.pdf"/>
    <dc:rights>Attribution 4.0 International</dc:rights>
  </rdf:Description>
</rdf:RDF>
Interner Vermerk
xmlui.Submission.submit.DescribeStep.inputForms.label.kops_note_fromSubmitter
Kontakt
URL der Originalveröffentl.
Prüfdatum der URL
Prüfungsdatum der Dissertation
Finanzierungsart
Kommentar zur Publikation
Allianzlizenz
Corresponding Authors der Uni Konstanz vorhanden
Internationale Co-Autor:innen
Universitätsbibliographie
Ja
Begutachtet
Unbekannt
Diese Publikation teilen