KonVid-150k : A Dataset for No-Reference Video Quality Assessment of Videos in-the-Wild

Lade...
Vorschaubild
Dateien
Goetz-Hahn_2-1t0bmybzek49b5.pdf
Goetz-Hahn_2-1t0bmybzek49b5.pdfGröße: 2.77 MBDownloads: 220
Datum
2021
Herausgeber:innen
Kontakt
ISSN der Zeitschrift
Electronic ISSN
ISBN
Bibliografische Daten
Verlag
Schriftenreihe
Auflagebezeichnung
ArXiv-ID
Internationale Patentnummer
Link zur Lizenz
Angaben zur Forschungsförderung
Projekt
Open Access-Veröffentlichung
Open Access Gold
Core Facility der Universität Konstanz
Gesperrt bis
Titel in einer weiteren Sprache
Forschungsvorhaben
Organisationseinheiten
Zeitschriftenheft
Publikationstyp
Zeitschriftenartikel
Publikationsstatus
Published
Erschienen in
Zusammenfassung

Video quality assessment (VQA) methods focus on particular degradation types, usually artificially induced on a small set of reference videos. Hence, most traditional VQA methods under-perform in-the-wild. Deep learning approaches have had limited success due to the small size and diversity of existing VQA datasets, either artificial or authentically distorted. We introduce a new in-the-wild VQA dataset that is substantially larger and diverse: KonVid-150k. It consists of a coarsely annotated set of 153,841 videos having five quality ratings each, and 1,596 videos with a minimum of 89 ratings each. Additionally, we propose new efficient VQA approaches (MLSP-VQA) relying on multi-level spatially pooled deep-features (MLSP). They are exceptionally well suited for training at scale, compared to deep transfer learning approaches. Our best method, MLSP-VQA-FF, improves the Spearman rank-order correlation coefficient (SRCC) performance metric on the commonly used KoNViD-1k in-the-wild benchmark dataset to 0.82. It surpasses the best existing deep-learning model (0.80 SRCC) and hand-crafted feature-based method (0.78 SRCC). We further investigate how alternative approaches perform under different levels of label noise, and dataset size, showing that MLSP-VQA-FF is the overall best method for videos in-the-wild. Finally, we show that the MLSP-VQA models trained on KonVid-150k sets the new state-of-the-art for cross-test performance on KoNViD-1k and LIVE-Qualcomm with a 0.83 and 0.64 SRCC, respectively. For KoNViD-1k this inter-dataset testing outperforms intra-dataset experiments, showing excellent generalization.

Zusammenfassung in einer weiteren Sprache
Fachgebiet (DDC)
004 Informatik
Schlagwörter
Datasets, deep transfer learning, multi-level spatially-pooled features, video quality assessment, video quality dataset
Konferenz
Rezension
undefined / . - undefined, undefined
Zitieren
ISO 690GÖTZ-HAHN, Franz, Vlad HOSU, Hanhe LIN, Dietmar SAUPE, 2021. KonVid-150k : A Dataset for No-Reference Video Quality Assessment of Videos in-the-Wild. In: IEEE Access. IEEE. 2021, 9, pp. 72139-72160. eISSN 2169-3536. Available under: doi: 10.1109/ACCESS.2021.3077642
BibTex
@article{GotzHahn2021KonVi-53766,
  year={2021},
  doi={10.1109/ACCESS.2021.3077642},
  title={KonVid-150k : A Dataset for No-Reference Video Quality Assessment of Videos in-the-Wild},
  volume={9},
  journal={IEEE Access},
  pages={72139--72160},
  author={Götz-Hahn, Franz and Hosu, Vlad and Lin, Hanhe and Saupe, Dietmar}
}
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/123456789/53766">
    <dcterms:issued>2021</dcterms:issued>
    <dspace:isPartOfCollection rdf:resource="https://kops.uni-konstanz.de/server/rdf/resource/123456789/36"/>
    <dc:language>eng</dc:language>
    <dcterms:hasPart rdf:resource="https://kops.uni-konstanz.de/bitstream/123456789/53766/1/Goetz-Hahn_2-1t0bmybzek49b5.pdf"/>
    <bibo:uri rdf:resource="https://kops.uni-konstanz.de/handle/123456789/53766"/>
    <dc:creator>Saupe, Dietmar</dc:creator>
    <dc:contributor>Lin, Hanhe</dc:contributor>
    <dc:creator>Hosu, Vlad</dc:creator>
    <dspace:hasBitstream rdf:resource="https://kops.uni-konstanz.de/bitstream/123456789/53766/1/Goetz-Hahn_2-1t0bmybzek49b5.pdf"/>
    <dcterms:abstract xml:lang="eng">Video quality assessment (VQA) methods focus on particular degradation types, usually artificially induced on a small set of reference videos. Hence, most traditional VQA methods under-perform in-the-wild. Deep learning approaches have had limited success due to the small size and diversity of existing VQA datasets, either artificial or authentically distorted. We introduce a new in-the-wild VQA dataset that is substantially larger and diverse: KonVid-150k. It consists of a coarsely annotated set of 153,841 videos having five quality ratings each, and 1,596 videos with a minimum of 89 ratings each. Additionally, we propose new efficient VQA approaches (MLSP-VQA) relying on multi-level spatially pooled deep-features (MLSP). They are exceptionally well suited for training at scale, compared to deep transfer learning approaches. Our best method, MLSP-VQA-FF, improves the Spearman rank-order correlation coefficient (SRCC) performance metric on the commonly used KoNViD-1k in-the-wild benchmark dataset to 0.82. It surpasses the best existing deep-learning model (0.80 SRCC) and hand-crafted feature-based method (0.78 SRCC). We further investigate how alternative approaches perform under different levels of label noise, and dataset size, showing that MLSP-VQA-FF is the overall best method for videos in-the-wild. Finally, we show that the MLSP-VQA models trained on KonVid-150k sets the new state-of-the-art for cross-test performance on KoNViD-1k and LIVE-Qualcomm with a 0.83 and 0.64 SRCC, respectively. For KoNViD-1k this inter-dataset testing outperforms intra-dataset experiments, showing excellent generalization.</dcterms:abstract>
    <dcterms:rights rdf:resource="http://creativecommons.org/licenses/by/4.0/"/>
    <dcterms:title>KonVid-150k : A Dataset for No-Reference Video Quality Assessment of Videos in-the-Wild</dcterms:title>
    <void:sparqlEndpoint rdf:resource="http://localhost/fuseki/dspace/sparql"/>
    <dcterms:available rdf:datatype="http://www.w3.org/2001/XMLSchema#dateTime">2021-05-27T09:36:18Z</dcterms:available>
    <dc:creator>Götz-Hahn, Franz</dc:creator>
    <dcterms:isPartOf rdf:resource="https://kops.uni-konstanz.de/server/rdf/resource/123456789/36"/>
    <dc:contributor>Hosu, Vlad</dc:contributor>
    <dc:contributor>Götz-Hahn, Franz</dc:contributor>
    <dc:rights>Attribution 4.0 International</dc:rights>
    <dc:contributor>Saupe, Dietmar</dc:contributor>
    <foaf:homepage rdf:resource="http://localhost:8080/"/>
    <dc:creator>Lin, Hanhe</dc:creator>
    <dc:date rdf:datatype="http://www.w3.org/2001/XMLSchema#dateTime">2021-05-27T09:36:18Z</dc:date>
  </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
Ja
Diese Publikation teilen