Skip to content
2000
Volume 18, Issue 9
  • ISSN: 2352-0965
  • E-ISSN: 2352-0973

Abstract

Introduction

Deep learning models often face challenges in achieving optimal accuracy when classifying multiclass datasets, particularly when the dataset size is limited. This study introduces Contrast Based Learning (CBL), a novel data augmentation technique designed to address data scarcity.

Methods

CBL innovatively concatenates multiple images and uses contrast learning to generate enriched datasets that exhibit a higher diversity of complex features. By focusing on the contrasts between various images, this method enhances the model's ability to learn nuanced features, thereby improving generalization and reducing overfitting.

Results

Unlike traditional data augmentation methods, which rely on basic transformations, CBL dynamically concatenates images from different classes, creating complex inputs that provides the model with a more comprehensive training dataset. Experimental results show that CBL significantly improves classification accuracy and outperforms state-of-the-art methods across multiple small-scale multiclass datasets.

Conclusion

The findings highlight the robustness of CBL in addressing data limitations, demonstrating its potential to advance the classification performance of deep learning models.

© 2025 Bentham Science Publishers
Loading

Article metrics loading...

/content/journals/raeeng/10.2174/0123520965298906241023110601
2024-12-26
2026-01-09

  • From This Site

    /content/journals/raeeng/10.2174/0123520965298906241023110601
    dcterms_title,dcterms_subject,pub_keyword
    -contentType:Contributor -contentType:Concept -contentType:Institution
    10
    5
Loading full text...

Full text loading...

References

  1. BrigatoL. IocchiL. A Close Look at Deep Learning with Small Data.ICPR202110.1109/ICPR48806.2021.9412492
     [Google Scholar]
  2. TakahashiR. MatsubaraT. UeharaK. Data augmentation using random image cropping and patching for deep CNNs.IEEE Trans. Circ. Syst. Video Tech.20203092917293110.1109/TCSVT.2019.2935128
     [Google Scholar]
  3. EngstromL. TranB. TsiprasD. SchmidtL. MadryA. A rotation and a translation suffice: Fooling cnns with simple transformations.ICLR2018
     [Google Scholar]
  4. BrostowG.J. FauqueurJ. CipollaR. Semantic object classes in video: A high-definition ground truth database.Pattern Recognit. Lett.2009302889710.1016/j.patrec.2008.04.005
     [Google Scholar]
  5. UllahM.B. CPU based YOLO: A real time object detection algorithm.ICLR202010.1109/TENSYMP50017.2020.9230778
     [Google Scholar]
  6. GirshickR. DonahueJ. DarrellT. MalikJ. Rich feature hierarchies for accurate object detection and semantic segmentationProceedings of the IEEE conference on computer vision and pattern recognition, 2014, pp. 580-587.10.1109/CVPR.2014.81
     [Google Scholar]
  7. BrigatoL. BarzB. IocchiL. DenzlerJ. Tune it or don’t use it: Benchmarking data-efficient image classificationProceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 1071-1080.10.1109/ICCVW54120.2021.00125
     [Google Scholar]
  8. AlzubaidiL. BaiJ. Al-SabaawiA. SantamaríaJ. AlbahriA.S. Al-dabbaghB.S.N. FadhelM.A. ManoufaliM. ZhangJ. Al-TimemyA.H. DuanY. AbdullahA. FarhanL. LuY. GuptaA. AlbuF. AbboshA. GuY. A survey on deep learning tools dealing with data scarcity: Definitions, challenges, solutions, tips, and applications.J. Big Data20231014610.1186/s40537‑023‑00727‑2
     [Google Scholar]
  9. LiH. LiJ. GuanX. LiangB. LaiY. LuoX. Research on Overfitting of Deep Learning2019 15th International Conference on Computational Intelligence and Security (CIS), 2019 Macao, China, 2019, pp. 78-81.10.1109/CIS.2019.00025
     [Google Scholar]
  10. HanM. LiA. GaoZ. MuD. LiuS. A survey of multi-class imbalanced data classification methods.J. Intell. Fuzzy Syst.20234422471250110.3233/JIFS‑221902
     [Google Scholar]
  11. WeissK. KhoshgoftaarT.M. WangD. A survey of transfer learning.J. Big Data201631910.1186/s40537‑016‑0043‑6
     [Google Scholar]
  12. DengJ. DongW. SocherR. LiL-J. LiK. Fei-FeiL. ImageNet: A large-scale hierarchical image database2009 IEEE Conference on Computer Vision and Pattern Recognition, Miami, FL, USA, 2009, pp. 248-255.10.1109/CVPR.2009.5206848
     [Google Scholar]
  13. ChenT. KornblithS. SwerskyK. NorouziM. HintonG.E. Big self-supervised models are strong semi-supervised learners.Adv. Neural Inf. Process. Syst.2020332224322255
     [Google Scholar]
  14. TerranceV. GrahamW.T. Dataset augmentation in feature spaceProceedings of the International Conference on Machine Learning (ICML), Workshop Track, 2017
     [Google Scholar]
  15. GengM. XuK. DingB. WangH. ZhangL. Learning data augmentation policies using augmented random searcharXiv preprint arXiv:1811.047682018201804768
     [Google Scholar]
  16. InoueH. Data augmentation by pairing samples for images classificationarXiv preprint arXiv:1801.029292018201802929
     [Google Scholar]
  17. SobujS.I. Leveraging Pre-trained CNNs for Efficient Feature Extraction in Rice Leaf Disease Classification2024 International Conference on Advances in Computing, Communication, Electrical, and Smart Systems (iCACCESS), 2024.10.1109/iCACCESS61735.2024.10499603
     [Google Scholar]
  18. KalpanaR. BennetM.A. An Image Feature Extraction Approach Using A Randomized Deep Neural Network Algorithm2023 4th International Conference on Computation, Automation and Knowledge Management (ICCAKM), 2023.10.1109/ICCAKM58659.2023.10449525
     [Google Scholar]
  19. YangZ. Enhancing Convolutional Neural Networks via separately trained kernels for digit recognition.Applied and Computational Engineering202454125425710.54254/2755‑2721/54/20241659
     [Google Scholar]
  20. AbdulkareemI.M. Proposed Approach for Object Detection and Recognition by Deep Learning Models Using Data AugmentationiJOE20242005314310.3991/ijoe.v20i05.47171
     [Google Scholar]
  21. MumuniA. MumuniF. Data augmentation: A comprehensive survey of modern approaches.Array20221610025810.1016/j.array.2022.100258
     [Google Scholar]
  22. ShortenC. KhoshgoftaarT.M. A survey on image data augmentation for deep learning.J. Big Data2019616010.1186/s40537‑019‑0197‑0
     [Google Scholar]
  23. GarceaF. SerraA. LambertiF. MorraL. Data augmentation for medical imaging: A systematic literature review.Comput. Biol. Med.202315210639110.1016/j.compbiomed.2022.106391 36549032
     [Google Scholar]
  24. BarzB. DenzlerJ. Deep learning on small datasets without pre-training using cosine lossProceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2020. pp. 1371-1380.10.1109/WACV45572.2020.9093286
     [Google Scholar]
  25. KimB. KimC. LeeJ. SongJ. ParkG. Data-Efficient Deep Learning Method for Image Classification Using Data Augmentation, Focal Cosine Loss, and EnsemblearXiv preprint2020
     [Google Scholar]
  26. LiangD. YangF. ZhangT. YangP. Understanding mixup training methods.IEEE Access20186587745878310.1109/ACCESS.2018.2872698
     [Google Scholar]
  27. ZhangH. CisseM. DauphinY.N. Lopez-PazD. mixup: Beyond empirical risk minimizationarXiv preprint2017201709412
     [Google Scholar]
  28. YunS. HanD. OhS.J. ChunS. ChoeJ. YooY. Cutmix: Regularization strategy to train strong classifiers with localizable featuresProceedings of the IEEE/CVF International Conference on Computer Vision, 2019, pp. 6023-6032.10.1109/ICCV.2019.00612
     [Google Scholar]
  29. ZhongZ. ZhengL. KangG. LiS. YangY. Random erasing data augmentationProceedings of the AAAI Conference on Artificial Intelligence, 2020, pp. 13001-13008.
     [Google Scholar]
  30. CubukE.D. ZophB. ManeD. VasudevanV. LeQ.V. Autoaugment: Learning augmentation policies from data arXiv preprint2018201809501
     [Google Scholar]
  31. BarretZ. Le QuocV. Neural architecture search with reinforcement learningInternational Conference on Learning Representatoins2017
     [Google Scholar]
  32. SuttonR.S. BartoA.G. The reinforcement learning problem. Reinforcement learning: An introduction.MIT Press Cambridge, MA19985185
     [Google Scholar]
  33. MinhT.N. SinnM. LamH.T. WistubaM. Automated image data preprocessing with deep reinforcement learningarXiv preprint2018201805886
     [Google Scholar]
  34. SalimansT. HoJ. ChenX. SidorS. SutskeverI. Evolution strategies as a scalable alternative to reinforcement learningarXiv preprint2017201703864
     [Google Scholar]
  35. ManiaH. GuyA. RechtB. Simple random search provides a competitive approach to reinforcement learningarXiv preprint2018201807055
     [Google Scholar]
  36. Ud DinN.M. Ul SabhaS. BhatM.R. AssadA. Axillary Lymph Node Metastasis prediction Using Deep Reinforcement Learning on Primary Tumor Biopsy Slides2023 3rd International conference on Artificial Intelligence and Signal Processing (AISP), VIJAYAWADA, 2023, pp. 1-5.10.1109/AISP57993.2023.10135039
     [Google Scholar]
  37. LiY. HuG. WangY. HospedalesT. RobertsonN.M. YangY. DADA: Differentiable Automatic Data AugmentationAvailable from: https://github.com/VDIGPKU/DADA
  38. ZhangX. WangZ. LiuD. LingQ. Dada: Deep adversarial data augmentation for extremely low data regime classification. Icassp 2019-2019 IEEE International Conference on Acoustics, Speech and Signal Processing, 2019, pp. 2807-2811.10.1109/ICASSP.2019.8683197
     [Google Scholar]
  39. RosensteinM.T. MarxZ. KaelblingL.P. DietterichT.G. To transfer or not to transferNIPS 2005 Workshop on Transfer learning, 2005.
     [Google Scholar]
  40. DayO. KhoshgoftaarT.M. A survey on heterogeneous transfer learning.J. Big Data2017412910.1186/s40537‑017‑0089‑0
     [Google Scholar]
  41. TanC. SunF. KongT. ZhangW. YangC. LiuC. A survey on deep transfer learning27th International Conference on Artificial Neural Networks Rhodes, Greece, October 4-7, 2018, pp.270-279.
     [Google Scholar]
  42. AyanaG. ParkJ. JeongJ.W. ChoeS. A novel multistage transfer learning for ultrasound breast cancer image classification.Diagnostics202212113510.3390/diagnostics12010135 35054303
     [Google Scholar]
  43. BorneaM. PanL. RosenthalS. FlorianR. SilA. Multilingual transfer learning for QA using translation as data augmentationProceedings of the AAAI Conference on Artificial Intelligence, 2021, pp. 12583-12591.10.1609/aaai.v35i14.17491
     [Google Scholar]
  44. BaurC. AlbarqouniS. NavabN. MelanoGANs: High resolution skin lesion synthesis with GANsarXiv preprint2018201804338
     [Google Scholar]
  45. ZhaoC. XiangS. WangY. CaiZ. ShenJ. ZhouS. ZhaoD. SuW. GuoS. LiS. Context-aware network fusing transformer and V-Net for semi-supervised segmentation of 3D left atrium.Expert Syst. Appl.202321411910510.1016/j.eswa.2022.119105
     [Google Scholar]
  46. XiangS. LiN. WangY. ZhouS. WeiJ. LiS. Automatic Delineation of the 3D Left Atrium From LGE-MRI: Actor-Critic Based Detection and Semi-Supervised Segmentation.IEEE J. Biomed. Health Inform.20242863545355610.1109/JBHI.2024.3373127 38442054
     [Google Scholar]
  47. KrizhevskyA. SutskeverI. HintonG.E. Imagenet classification with deep convolutional neural networksAdv Neural Inf Process Syst
     [Google Scholar]
  48. ZhangX. WangZ. LiuD. LinQ. LingQ. Deep Adversarial Data Augmentation for Extremely Low Data Regimes.IEEE Trans. Circ. Syst. Video Tech.2021311152810.1109/TCSVT.2020.2967419
     [Google Scholar]
  49. ArshadT. JunpingZ. WangQ. Multiclass Classification of Remote Sensing Images Using Deep Learning TechniquesIGARSS 2023 - 2023 IEEE International Geoscience and Remote Sensing Symposium, 2023, pp. 7234–7237.10.1109/IGARSS52108.2023.10283456
     [Google Scholar]
/content/journals/raeeng/10.2174/0123520965298906241023110601
Loading
Efficient Multiclass Classification of Small Datasets Using a Novel Contrast Based Learning Approach
Recent Advances in Electrical & Electronic Engineering 18, 1678 (2025); https://doi.org/10.2174/0123520965298906241023110601
/content/journals/raeeng/10.2174/0123520965298906241023110601
/content/journals/raeeng/10.2174/0123520965298906241023110601
Loading

Data & Media loading...


  • Article Type:     Research Article
Keyword(s): accuracy; Artificial Intelligence; convolutional neural network; data augmentation; deep learning; overfitting

Most Cited Most Cited RSS feed

This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error
Please enter a valid_number test