A semi-supervised network based on feature embeddings for image classification

被引:1
|
作者
Nuhoho, Raphael Elimeli [1 ]
Chen Wenyu [1 ]
Baffour, Adu Asare [2 ]
机构
[1] Univ Elect Sci & Technol China, Computat Intelligence Lab, Sch Comp Sci & Technol, 2006 Xiyuan Ave, Chengdu 611731, Sichuan, Peoples R China
[2] Univ Elect Sci & Technol China, Sch Informat & Software Engn, Chengdu, Peoples R China
关键词
clustering; convolutional neural network; feature embedding; semi-supervised learning;
D O I
10.1111/exsy.12908
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Deep learning approaches, including convolutional neural networks, are suitable for image classification tasks with well-labelled data. Unfortunately, we do not always have sufficiently labelled data. Recent methods attempt to leverage labelled and unlabelled data using fine-tuning or transfer learning. However, these methods rely on low-level image features. This article departs from recent works and proposes a new semi-supervised learning network that constitutes a convolutional branch and a neighbour cluster branch. Also, we introduce a new loss function that carefully optimizes the network according to the labelled/unlabelled data. In this way, we reduce any tendency to rely on low-level features, which is the case in current methods. We use datasets from three different domains (hand-written digits, natural images, and objects) to analyse the performance of our method. Experimental analysis shows that the network performs better by learning inherent discrimination features when integrating unlabelled data into the model's training process. Our proposed approach also provides strong generalization in the context of transfer learning. Finally, this study shows that the proposed loss function optimizes the network to produce more efficient feature embeddings for domain adaptation.
引用
收藏
页数:16
相关论文
共 50 条
  • [41] Adaptive Feature Selection and Feature Fusion for Semi-supervised Classification
    Wei Du
    Ronald Phlypo
    Tülay Adalı
    [J]. Journal of Signal Processing Systems, 2019, 91 : 521 - 537
  • [42] Semi-supervised Image Labelling Using Barycentric Graph Embeddings
    Robles-Kelly, Antonio
    Wei, Ran
    [J]. 2016 23RD INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION (ICPR), 2016, : 1518 - 1523
  • [43] Semi-supervised tensor learning for image classification
    Jianguang Zhang
    Yahong Han
    Jianmin Jiang
    [J]. Multimedia Systems, 2017, 23 : 63 - 73
  • [44] Semi-supervised tensor learning for image classification
    Zhang, Jianguang
    Han, Yahong
    Jiang, Jianmin
    [J]. MULTIMEDIA SYSTEMS, 2017, 23 (01) : 63 - 73
  • [45] Semi-supervised Learning for Image Modality Classification
    de Herrera, Alba Garcia Seco
    Markonis, Dimitrios
    Joyseeree, Ranveer
    Schaer, Roger
    Foncubierta-Rodriguez, Antonio
    Mueller, Henning
    [J]. MULTIMODAL RETRIEVAL IN THE MEDICAL DOMAIN, MRMD 2015, 2015, 9059 : 85 - 98
  • [46] Semi-supervised image classification in likelihood space
    Duan, Rong
    Jiang, Wei
    Man, Hong
    [J]. 2006 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, ICIP 2006, PROCEEDINGS, 2006, : 957 - +
  • [47] Semi-Supervised Classification of Wafer Map Based on Ladder Network
    Kong, Yuting
    Ni, Dong
    [J]. 2018 14TH IEEE INTERNATIONAL CONFERENCE ON SOLID-STATE AND INTEGRATED CIRCUIT TECHNOLOGY (ICSICT), 2018, : 723 - 726
  • [48] GMDH-based semi-supervised feature selection for customer classification
    Xiao, Jin
    Cao, Hanwen
    Jiang, Xiaoyi
    Gu, Xin
    Xie, Ling
    [J]. KNOWLEDGE-BASED SYSTEMS, 2017, 132 : 236 - 248
  • [49] Mass Classification in Mammogram with Semi-Supervised Relief Based Feature Selection
    Liu, Xiaoming
    Liu, Jun
    Feng, Zhilin
    Xu, Xin
    Tang, J.
    [J]. FIFTH INTERNATIONAL CONFERENCE ON GRAPHIC AND IMAGE PROCESSING (ICGIP 2013), 2014, 9069
  • [50] Multimodal semi-supervised learning for image classification
    Guillaumin, Matthieu
    Verbeek, Jakob
    Schmid, Cordelia
    [J]. 2010 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2010, : 902 - 909