We propose a novel semi-supervised learning method to leverage unlabeled data alongside minimal annotated data and improve medical imaging classification performance in realistic scenarios with limited labeling budgets to afford data annotations. Our method introduces distance correlation to minimize correlations between feature representations from different views of the same image encoded with non-coupled deep neural networks architectures. In addition, it incorporates a data-driven graph-attention based regularization strategy to model affinities among images within the unlabeled data by exploiting their inherent relational information in the feature space. We conduct extensive experiments on four medical imaging benchmark data sets involving X-ray, dermoscopic, magnetic resonance, and computer tomography imaging on single and multi-label medical imaging classification scenarios. Our experiments demonstrate the effectiveness of our method in achieving very competitive performance and outperforming several state-of-the-art semi-supervised learning methods. Furthermore, they confirm the suitability of distance correlation as a versatile dependence measure and the benefits of the proposed graph-attention based regularization for semi-supervised learning in medical imaging analysis.
Diaz Berenguer, A , Kvasnytsia, M , Bossa Bossa, MN , Mukherjee, T , Deligiannis, N & Sahli, H 2024, ' Semi-supervised medical image classification via distance correlation minimization and graph attention regularization ', Medical Image Analysis , vol. 94, 103107, pp. 1-16.
Diaz Berenguer, A. , Kvasnytsia, M. , Bossa Bossa, M. N. , Mukherjee, T. , Deligiannis, N. , & Sahli, H. (2024). Semi-supervised medical image classification via distance correlation minimization and graph attention regularization . Medical Image Analysis , 94 , 1-16. [103107].
@article{c5f18eca31f0457a875898ba29082a92,
title = " Semi-supervised medical image classification via distance correlation minimization and graph attention regularization " ,
abstract = " We propose a novel semi-supervised learning method to leverage unlabeled data alongside minimal annotated data and improve medical imaging classification performance in realistic scenarios with limited labeling budgets to afford data annotations. Our method introduces distance correlation to minimize correlations between feature representations from different views of the same image encoded with non-coupled deep neural networks architectures. In addition, it incorporates a data-driven graph-attention based regularization strategy to model affinities among images within the unlabeled data by exploiting their inherent relational information in the feature space. We conduct extensive experiments on four medical imaging benchmark data sets involving X-ray, dermoscopic, magnetic resonance, and computer tomography imaging on single and multi-label medical imaging classification scenarios. Our experiments demonstrate the effectiveness of our method in achieving very competitive performance and outperforming several state-of-the-art semi-supervised learning methods. Furthermore, they confirm the suitability of distance correlation as a versatile dependence measure and the benefits of the proposed graph-attention based regularization for semi-supervised learning in medical imaging analysis. " ,
keywords = " Deep neural networks, Distance correlation, Graph attention, Medical image classification, Semi-supervised learning " ,
author = " {Diaz Berenguer}, Abel and Maryna Kvasnytsia and {Bossa Bossa}, {Mat{'i}as Nicol{'a}s} and Tanmoy Mukherjee and Nikos Deligiannis and Hichem Sahli " ,
note = " Funding Information: We want to thank Wang et al. (2017), Codella et al. (2019), Bien et al. (2018) , and Kollias et al. (2022) for providing access to extensive and high-quality data. We also want to thank the VUB-HPC team for providing access to the services of the Cluster Hydra supporting the development of this work. This research received funding from the INNOVIRIS project “ANTICIPATE – AugmeNTed IntelligenCe In orthopaedics TrEatments” (grant ID: BHG/2020-RDIR-6a ). Publisher Copyright: { extcopyright} 2024 " ,
year = " 2024 " ,
month = may,
doi = " https://doi.org/10.1016/j.media.2024.103107 " ,
language = " English " ,
volume = " 94 " ,
pages = " 116 " ,
journal = " Medical Image Analysis " ,
issn = " 1361-8415 " ,
publisher = " Elsevier " ,
}