Academic Journal
View in EDS
Linked Full Text

Deep one-class probability learning for end-to-end image classification.

Bibliographic Details
Title: Deep one-class probability learning for end-to-end image classification.
Authors: Liu J; School of Computer Science and Engineering, Nanjing University of Science and Technology, Nanjing, 210094, China., Zhang W; School of Computer Science and Engineering, Nanjing University of Science and Technology, Nanjing, 210094, China. Electronic address: whzhang@njust.edu.cn., Liu F; School of Computer Science and Engineering, Nanjing University of Science and Technology, Nanjing, 210094, China., Yang J; School of Computer Science and Engineering, Nanjing University of Science and Technology, Nanjing, 210094, China., Xiao L; School of Computer Science and Engineering, Nanjing University of Science and Technology, Nanjing, 210094, China.
Source: Neural networks : the official journal of the International Neural Network Society [Neural Netw] 2025 May; Vol. 185, pp. 107201. Date of Electronic Publication: 2025 Jan 27.
Publication Type: Journal Article
Language: English
Journal Info: 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-
MeSH Terms: Image Processing, Computer-Assisted*/methods , Deep Learning* , Probability Learning* , Neural Networks, Computer*, Algorithms ; Humans ; Probability
Abstract: One-class learning has many application potentials in novelty, anomaly, and outlier detection systems. It aims to distinguish both positive and negative samples with a model trained via only positive samples or one-class annotated samples. With the difficulty in training an end-to-end classification network, existing methods usually make decisions indirectly. To fully exploit the learning capability of a deep network, in this paper, we propose to design a deep end-to-end binary image classifier based on convolutional neural network with input of image and output of classification result. Without negative training samples, we establish a probabilistic model driven by an energy to learn the distribution of positive samples. The energy is proposed based on the output of the network which subtly models the deep discriminations into statistics. During optimization, to overcome the difficulty of distribution estimation, we propose a novel particle swarm optimization algorithm based sampling method. Compared with existing methods, the proposed method is able to directly output classification results without additional thresholding or estimating operations. Moreover, the deep network is directly optimized via the probabilistic model which results in better adaptation of positive distribution and classification task. Experiments demonstrate the effectiveness and state-of-the-art performance of the proposed method.
(Copyright © 2025 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: Deep neural network; Image classification; One-class learning; Probabilistic model
Entry Date(s): Date Created: 20250204 Date Completed: 20250505 Latest Revision: 20250525
Update Code: 20260130
DOI: 10.1016/j.neunet.2025.107201
PMID: 39903959
Database: MEDLINE
FullText Links:
  – Type: other
    Url: https://resolver.ebsco.com:443/public/rma-ftfapi/ejs/direct?AccessToken=48EBA89E6119693CBB20&Show=Object
Text:
  Availability: 0
CustomLinks:
  – Url: https://www.doi.org/10.1016/j.neunet.2025.107201?
    Name: ScienceDirect (all content) (s7799221)
    Category: fullText
    Text: View record from ScienceDirect
    MouseOverText: View record from ScienceDirect
Header DbId: cmedm
DbLabel: MEDLINE
An: 39903959
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Deep one-class probability learning for end-to-end image classification.
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AU" term="%22Liu+J%22">Liu J</searchLink>; School of Computer Science and Engineering, Nanjing University of Science and Technology, Nanjing, 210094, China.<br /><searchLink fieldCode="AU" term="%22Zhang+W%22">Zhang W</searchLink>; School of Computer Science and Engineering, Nanjing University of Science and Technology, Nanjing, 210094, China. Electronic address: whzhang@njust.edu.cn.<br /><searchLink fieldCode="AU" term="%22Liu+F%22">Liu F</searchLink>; School of Computer Science and Engineering, Nanjing University of Science and Technology, Nanjing, 210094, China.<br /><searchLink fieldCode="AU" term="%22Yang+J%22">Yang J</searchLink>; School of Computer Science and Engineering, Nanjing University of Science and Technology, Nanjing, 210094, China.<br /><searchLink fieldCode="AU" term="%22Xiao+L%22">Xiao L</searchLink>; School of Computer Science and Engineering, Nanjing University of Science and Technology, Nanjing, 210094, China.
– 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] 2025 May; Vol. 185, pp. 107201. <i>Date of Electronic Publication: </i>2025 Jan 27.
– 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="%22Deep+Learning%22">Deep Learning*</searchLink> <br /><searchLink fieldCode="MM" term="%22Probability+Learning%22">Probability Learning*</searchLink> <br /><searchLink fieldCode="MM" term="%22Neural+Networks%2C+Computer%22">Neural Networks, Computer*</searchLink><br /><searchLink fieldCode="MH" term="%22Algorithms%22">Algorithms</searchLink> ; <searchLink fieldCode="MH" term="%22Humans%22">Humans</searchLink> ; <searchLink fieldCode="MH" term="%22Probability%22">Probability</searchLink>
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: One-class learning has many application potentials in novelty, anomaly, and outlier detection systems. It aims to distinguish both positive and negative samples with a model trained via only positive samples or one-class annotated samples. With the difficulty in training an end-to-end classification network, existing methods usually make decisions indirectly. To fully exploit the learning capability of a deep network, in this paper, we propose to design a deep end-to-end binary image classifier based on convolutional neural network with input of image and output of classification result. Without negative training samples, we establish a probabilistic model driven by an energy to learn the distribution of positive samples. The energy is proposed based on the output of the network which subtly models the deep discriminations into statistics. During optimization, to overcome the difficulty of distribution estimation, we propose a novel particle swarm optimization algorithm based sampling method. Compared with existing methods, the proposed method is able to directly output classification results without additional thresholding or estimating operations. Moreover, the deep network is directly optimized via the probabilistic model which results in better adaptation of positive distribution and classification task. Experiments demonstrate the effectiveness and state-of-the-art performance of the proposed method.<br /> (Copyright © 2025 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>Deep neural network; Image classification; One-class learning; Probabilistic model
– Name: DateEntry
  Label: Entry Date(s)
  Group: Date
  Data: <i>Date Created: </i>20250204 <i>Date Completed: </i>20250505 <i>Latest Revision: </i>20250525
– Name: DateUpdate
  Label: Update Code
  Group: Date
  Data: 20260130
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.1016/j.neunet.2025.107201
– Name: AN
  Label: PMID
  Group: ID
  Data: 39903959
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=cmedm&AN=39903959
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1016/j.neunet.2025.107201
    Languages:
      – Code: eng
        Text: English
    PhysicalDescription:
      Pagination:
        StartPage: 107201
    Subjects:
      – SubjectFull: Algorithms
        Type: general
      – SubjectFull: Humans
        Type: general
      – SubjectFull: Probability
        Type: general
      – SubjectFull: Image Processing, Computer-Assisted methods
        Type: general
      – SubjectFull: Deep Learning
        Type: general
      – SubjectFull: Probability Learning
        Type: general
      – SubjectFull: Neural Networks, Computer
        Type: general
    Titles:
      – TitleFull: Deep one-class probability learning for end-to-end image classification.
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Liu J
      – PersonEntity:
          Name:
            NameFull: Zhang W
      – PersonEntity:
          Name:
            NameFull: Liu F
      – PersonEntity:
          Name:
            NameFull: Yang J
      – PersonEntity:
          Name:
            NameFull: Xiao L
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 05
              Text: 2025 May
              Type: published
              Y: 2025
          Identifiers:
            – Type: issn-electronic
              Value: 1879-2782
          Numbering:
            – Type: volume
              Value: 185
          Titles:
            – TitleFull: Neural networks : the official journal of the International Neural Network Society
              Type: main
ResultId 1