Academic Journal
Προβολή σε EDS
Συνδεδεμένο Πλήρες Κείμενο

On the inherent robustness of one-stage object detection against out-of-distribution data.

Λεπτομέρειες βιβλιογραφικής εγγραφής
Τίτλος: On the inherent robustness of one-stage object detection against out-of-distribution data.
Συγγραφείς: Martinez-Seras A; IDEKO, Basque Research and Technology Alliance (BRTA), Elgoibar, 20870, Spain. Electronic address: amartinezseras@ideko.es., Del Ser J; TECNALIA, Basque Research and Technology Alliance (BRTA), Derio, 48160, Spain; University of the Basque Country (UPV/EHU), Leioa, 48940, Spain. Electronic address: javier.delser@tecnalia.com., Olivares-Rad A; University of the Basque Country (UPV/EHU), Leioa, 48940, Spain., Andres A; TECNALIA, Basque Research and Technology Alliance (BRTA), Derio, 48160, Spain; University of Deusto, Bilbao, 48007, Spain., Garcia-Bringas P; University of Deusto, Bilbao, 48007, Spain.
Πηγή: Neural networks : the official journal of the International Neural Network Society [Neural Netw] 2026 Jul; Vol. 199, pp. 108683. Date of Electronic Publication: 2026 Feb 04.
Τύπος έκδοσης: Journal Article
Γλώσσα: English
Στοιχεία περιοδικού: Publisher: Pergamon Press Country of Publication: United States NLM ID: 8805018 Publication Model: Print-Electronic Cited Medium: Internet ISSN: 1879-2782 (Electronic) Linking ISSN: 08936080 NLM ISO Abbreviation: Neural Netw Subsets: MEDLINE
Imprint Name(s): Original Publication: New York : Pergamon Press, [c1988-
Ιατρικοί όροι (MeSH): Image Processing, Computer-Assisted*/methods , Detection Algorithms*, Algorithms ; Humans
Περίληψη: Robustness is a fundamental aspect for developing safe and trustworthy models, particularly when they are deployed in the open world. In this work we analyze the inherent capability of one-stage object detectors to robustly operate in the presence of out-of-distribution (OoD) data. Specifically, we propose a novel detection algorithm for detecting unknown objects in image data, which leverages the features extracted by the model from each sample. Differently from other recent approaches in the literature, our proposal does not require retraining the object detector, thereby allowing for the use of pretrained models. Our proposed OoD detector exploits the application of supervised dimensionality reduction techniques to mitigate the effects of the curse of dimensionality on the features extracted by the model. Furthermore, it utilizes high-resolution feature maps to identify potential unknown objects in an unsupervised fashion. Our experiments analyze the Pareto trade-off between the performance detecting known and unknown objects resulting from different algorithmic configurations and inference confidence thresholds. We also compare the performance of our proposed algorithm to that of logits-based post-hoc OoD methods, as well as possible fusion strategies. Finally, we discuss on the competitiveness of all tested methods against state-of-the-art OoD approaches for object detection models over the recently published Unknown Object Detection benchmark. The obtained results verify that the performance of avant-garde post-hoc OoD detectors can be further improved when combined with our proposed algorithm.
(Copyright © 2026 Elsevier Ltd. All rights reserved.)
Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Contributed Indexing: Keywords: Open-world object detection; Out-of-distribution detection; Safe artificial intelligence; Trustworthy artificial intelligence
Entry Date(s): Date Created: 20260212 Date Completed: 20260707 Latest Revision: 20260707
Update Code: 20260708
DOI: 10.1016/j.neunet.2026.108683
PMID: 41679045
Βάση Δεδομένων: MEDLINE
FullText Links:
  – Type: other
    Url: https://resolver.ebsco.com:443/public/rma-ftfapi/ejs/direct?AccessToken=4A4EB73AF450EBA42777&Show=Object
Text:
  Availability: 0
CustomLinks:
  – Url: https://www.doi.org/10.1016/j.neunet.2026.108683?
    Name: ScienceDirect (all content) (s7799221)
    Category: fullText
    Text: View record from ScienceDirect
    MouseOverText: View record from ScienceDirect
Header DbId: cmedm
DbLabel: MEDLINE
An: 41679045
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: On the inherent robustness of one-stage object detection against out-of-distribution data.
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AU" term="%22Martinez-Seras+A%22">Martinez-Seras A</searchLink>; IDEKO, Basque Research and Technology Alliance (BRTA), Elgoibar, 20870, Spain. Electronic address: amartinezseras@ideko.es.<br /><searchLink fieldCode="AU" term="%22Del+Ser+J%22">Del Ser J</searchLink>; TECNALIA, Basque Research and Technology Alliance (BRTA), Derio, 48160, Spain; University of the Basque Country (UPV/EHU), Leioa, 48940, Spain. Electronic address: javier.delser@tecnalia.com.<br /><searchLink fieldCode="AU" term="%22Olivares-Rad+A%22">Olivares-Rad A</searchLink>; University of the Basque Country (UPV/EHU), Leioa, 48940, Spain.<br /><searchLink fieldCode="AU" term="%22Andres+A%22">Andres A</searchLink>; TECNALIA, Basque Research and Technology Alliance (BRTA), Derio, 48160, Spain; University of Deusto, Bilbao, 48007, Spain.<br /><searchLink fieldCode="AU" term="%22Garcia-Bringas+P%22">Garcia-Bringas P</searchLink>; University of Deusto, Bilbao, 48007, Spain.
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="JN" term="%228805018%22">Neural networks : the official journal of the International Neural Network Society</searchLink> [Neural Netw] 2026 Jul; Vol. 199, pp. 108683. <i>Date of Electronic Publication: </i>2026 Feb 04.
– Name: TypePub
  Label: Publication Type
  Group: TypPub
  Data: Journal Article
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: TitleSource
  Label: Journal Info
  Group: Src
  Data: <i>Publisher: </i><searchLink fieldCode="PB" term="%22Pergamon+Press%22">Pergamon Press </searchLink><i>Country of Publication: </i>United States <i>NLM ID: </i>8805018 <i>Publication Model: </i>Print-Electronic <i>Cited Medium: </i>Internet <i>ISSN: </i>1879-2782 (Electronic) <i>Linking ISSN: </i><searchLink fieldCode="IS" term="%2208936080%22">08936080 </searchLink><i>NLM ISO Abbreviation: </i>Neural Netw <i>Subsets: </i>MEDLINE
– Name: PublisherInfo
  Label: Imprint Name(s)
  Group: PubInfo
  Data: <i>Original Publication</i>: New York : Pergamon Press, [c1988-
– Name: SubjectMESH
  Label: MeSH Terms
  Group: Su
  Data: <searchLink fieldCode="MM" term="%22Image+Processing%2C+Computer-Assisted%22">Image Processing, Computer-Assisted*</searchLink>/<searchLink fieldCode="MM" term="%22Image+Processing%2C+Computer-Assisted+methods%22">methods</searchLink> <br /><searchLink fieldCode="MM" term="%22Detection+Algorithms%22">Detection Algorithms*</searchLink><br /><searchLink fieldCode="MH" term="%22Algorithms%22">Algorithms</searchLink> ; <searchLink fieldCode="MH" term="%22Humans%22">Humans</searchLink>
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: Robustness is a fundamental aspect for developing safe and trustworthy models, particularly when they are deployed in the open world. In this work we analyze the inherent capability of one-stage object detectors to robustly operate in the presence of out-of-distribution (OoD) data. Specifically, we propose a novel detection algorithm for detecting unknown objects in image data, which leverages the features extracted by the model from each sample. Differently from other recent approaches in the literature, our proposal does not require retraining the object detector, thereby allowing for the use of pretrained models. Our proposed OoD detector exploits the application of supervised dimensionality reduction techniques to mitigate the effects of the curse of dimensionality on the features extracted by the model. Furthermore, it utilizes high-resolution feature maps to identify potential unknown objects in an unsupervised fashion. Our experiments analyze the Pareto trade-off between the performance detecting known and unknown objects resulting from different algorithmic configurations and inference confidence thresholds. We also compare the performance of our proposed algorithm to that of logits-based post-hoc OoD methods, as well as possible fusion strategies. Finally, we discuss on the competitiveness of all tested methods against state-of-the-art OoD approaches for object detection models over the recently published Unknown Object Detection benchmark. The obtained results verify that the performance of avant-garde post-hoc OoD detectors can be further improved when combined with our proposed algorithm.<br /> (Copyright © 2026 Elsevier Ltd. All rights reserved.)
– Name: Abstract
  Label: Competing Interests
  Group: Ab
  Data: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
– Name: SubjectMinor
  Label: Contributed Indexing
  Group:
  Data: <i>Keywords: </i>Open-world object detection; Out-of-distribution detection; Safe artificial intelligence; Trustworthy artificial intelligence
– Name: DateEntry
  Label: Entry Date(s)
  Group: Date
  Data: <i>Date Created: </i>20260212 <i>Date Completed: </i>20260707 <i>Latest Revision: </i>20260707
– Name: DateUpdate
  Label: Update Code
  Group: Date
  Data: 20260708
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.1016/j.neunet.2026.108683
– Name: AN
  Label: PMID
  Group: ID
  Data: 41679045
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=cmedm&AN=41679045
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1016/j.neunet.2026.108683
    Languages:
      – Code: eng
        Text: English
    PhysicalDescription:
      Pagination:
        StartPage: 108683
    Subjects:
      – SubjectFull: Algorithms
        Type: general
      – SubjectFull: Humans
        Type: general
      – SubjectFull: Image Processing, Computer-Assisted methods
        Type: general
      – SubjectFull: Detection Algorithms
        Type: general
    Titles:
      – TitleFull: On the inherent robustness of one-stage object detection against out-of-distribution data.
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Martinez-Seras A
      – PersonEntity:
          Name:
            NameFull: Del Ser J
      – PersonEntity:
          Name:
            NameFull: Olivares-Rad A
      – PersonEntity:
          Name:
            NameFull: Andres A
      – PersonEntity:
          Name:
            NameFull: Garcia-Bringas P
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 07
              Text: 2026 Jul
              Type: published
              Y: 2026
          Identifiers:
            – Type: issn-electronic
              Value: 1879-2782
          Numbering:
            – Type: volume
              Value: 199
          Titles:
            – TitleFull: Neural networks : the official journal of the International Neural Network Society
              Type: main
ResultId 1