Loading [MathJax]/extensions/tex2jax.js

Distribution-Aware Multi-Label FixMatch for Semi-Supervised Learning on CheXpert.

Publikation: Beitrag in Buch/Bericht/Sammelwerk/KonferenzbandAufsatz in KonferenzbandForschungPeer-Review

Details

OriginalspracheEnglisch
Titel des SammelwerksProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops
Herausgeber (Verlag)Institute of Electrical and Electronics Engineers Inc.
Seiten2295-2304
Seitenumfang10
ISBN (elektronisch)979-8-3503-6547-4
ISBN (Print)979-8-3503-6548-1
PublikationsstatusVeröffentlicht - 16 Juni 2024

Publikationsreihe

NameIEEE Computer Society Conference on Computer Vision and Pattern Recognition workshops
ISSN (Print)2160-7508
ISSN (elektronisch)2160-7516

Abstract

Semi-supervised learning (SSL) has achieved remarkable success for multiclass classification in recent years, yielding a promising solution for medical image classification where labeled data is scarce but unlabeled images are accessible. In the context of multi-label problems however, SSL is still under-explored. In this work we adapt Fix-Match to the multi-label scenario, specifically focusing on CheXpert, a multi-label chest X-ray classification dataset which is imbalanced and only partially labeled. Leveraging distribution alignment, our proposed method, ML-FixMatch+DA, achieves solid performance gains in SSL tasks (AUC: +2.6%) and in a missing label scenario (AUC: +1.9%). In contrast to previous work we achieve a performance gain on CheXpert using FixMatch. We show that in contrast to multiclass FixMatch, where distribution alignment is optional, it is essential for multi-label FixMatch to handle class imbalance and generate reliable (positive and negative) pseudo-labels. Our pseudo-label selection is based on a single threshold for all classes and handles imbalance with no prior knowledge on label distributions. Our adaptation keeps the simplicity of the original multi-class FixMatch with no added hyperparameters (even for imbalanced data) and demonstrates the feasibility of simple SSL for multi-label problems, filling a crucial gap in the literature.

ASJC Scopus Sachgebiete

Zitieren

Distribution-Aware Multi-Label FixMatch for Semi-Supervised Learning on CheXpert. / Ihler, Sontje; Kuhnke, Felix; Kuhlgatz, Timo et al.
Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops. Institute of Electrical and Electronics Engineers Inc., 2024. S. 2295-2304 (IEEE Computer Society Conference on Computer Vision and Pattern Recognition workshops).

Publikation: Beitrag in Buch/Bericht/Sammelwerk/KonferenzbandAufsatz in KonferenzbandForschungPeer-Review

Ihler, S, Kuhnke, F, Kuhlgatz, T & Seel, T 2024, Distribution-Aware Multi-Label FixMatch for Semi-Supervised Learning on CheXpert. in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops. IEEE Computer Society Conference on Computer Vision and Pattern Recognition workshops, Institute of Electrical and Electronics Engineers Inc., S. 2295-2304. https://doi.org/10.1109/CVPRW63382.2024.00235
Ihler, S., Kuhnke, F., Kuhlgatz, T., & Seel, T. (2024). Distribution-Aware Multi-Label FixMatch for Semi-Supervised Learning on CheXpert. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops (S. 2295-2304). (IEEE Computer Society Conference on Computer Vision and Pattern Recognition workshops). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/CVPRW63382.2024.00235
Ihler S, Kuhnke F, Kuhlgatz T, Seel T. Distribution-Aware Multi-Label FixMatch for Semi-Supervised Learning on CheXpert. in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops. Institute of Electrical and Electronics Engineers Inc. 2024. S. 2295-2304. (IEEE Computer Society Conference on Computer Vision and Pattern Recognition workshops). Epub 2024 Jun 1. doi: 10.1109/CVPRW63382.2024.00235
Ihler, Sontje ; Kuhnke, Felix ; Kuhlgatz, Timo et al. / Distribution-Aware Multi-Label FixMatch for Semi-Supervised Learning on CheXpert. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops. Institute of Electrical and Electronics Engineers Inc., 2024. S. 2295-2304 (IEEE Computer Society Conference on Computer Vision and Pattern Recognition workshops).
Download
@inproceedings{a80b16e0a99d48d6891b50ebdabcd4be,
title = "Distribution-Aware Multi-Label FixMatch for Semi-Supervised Learning on CheXpert.",
abstract = "Semi-supervised learning (SSL) has achieved remarkable success for multiclass classification in recent years, yielding a promising solution for medical image classification where labeled data is scarce but unlabeled images are accessible. In the context of multi-label problems however, SSL is still under-explored. In this work we adapt Fix-Match to the multi-label scenario, specifically focusing on CheXpert, a multi-label chest X-ray classification dataset which is imbalanced and only partially labeled. Leveraging distribution alignment, our proposed method, ML-FixMatch+DA, achieves solid performance gains in SSL tasks (AUC: +2.6%) and in a missing label scenario (AUC: +1.9%). In contrast to previous work we achieve a performance gain on CheXpert using FixMatch. We show that in contrast to multiclass FixMatch, where distribution alignment is optional, it is essential for multi-label FixMatch to handle class imbalance and generate reliable (positive and negative) pseudo-labels. Our pseudo-label selection is based on a single threshold for all classes and handles imbalance with no prior knowledge on label distributions. Our adaptation keeps the simplicity of the original multi-class FixMatch with no added hyperparameters (even for imbalanced data) and demonstrates the feasibility of simple SSL for multi-label problems, filling a crucial gap in the literature. ",
keywords = "semi supervised learning, CheXpert, multi-label classification, Missing Labels, X-ray classification, Semi-supervised multi-label learning, Semi-supervised learning, Imbalanced semi-supervised learning, Medical Image Analysis, Multi-Label Classification",
author = "Sontje Ihler and Felix Kuhnke and Timo Kuhlgatz and Thomas Seel",
year = "2024",
month = jun,
day = "16",
doi = "10.1109/CVPRW63382.2024.00235",
language = "English",
isbn = "979-8-3503-6548-1",
series = "IEEE Computer Society Conference on Computer Vision and Pattern Recognition workshops",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "2295--2304",
booktitle = "Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops",
address = "United States",

}

Download

TY - GEN

T1 - Distribution-Aware Multi-Label FixMatch for Semi-Supervised Learning on CheXpert.

AU - Ihler, Sontje

AU - Kuhnke, Felix

AU - Kuhlgatz, Timo

AU - Seel, Thomas

PY - 2024/6/16

Y1 - 2024/6/16

N2 - Semi-supervised learning (SSL) has achieved remarkable success for multiclass classification in recent years, yielding a promising solution for medical image classification where labeled data is scarce but unlabeled images are accessible. In the context of multi-label problems however, SSL is still under-explored. In this work we adapt Fix-Match to the multi-label scenario, specifically focusing on CheXpert, a multi-label chest X-ray classification dataset which is imbalanced and only partially labeled. Leveraging distribution alignment, our proposed method, ML-FixMatch+DA, achieves solid performance gains in SSL tasks (AUC: +2.6%) and in a missing label scenario (AUC: +1.9%). In contrast to previous work we achieve a performance gain on CheXpert using FixMatch. We show that in contrast to multiclass FixMatch, where distribution alignment is optional, it is essential for multi-label FixMatch to handle class imbalance and generate reliable (positive and negative) pseudo-labels. Our pseudo-label selection is based on a single threshold for all classes and handles imbalance with no prior knowledge on label distributions. Our adaptation keeps the simplicity of the original multi-class FixMatch with no added hyperparameters (even for imbalanced data) and demonstrates the feasibility of simple SSL for multi-label problems, filling a crucial gap in the literature.

AB - Semi-supervised learning (SSL) has achieved remarkable success for multiclass classification in recent years, yielding a promising solution for medical image classification where labeled data is scarce but unlabeled images are accessible. In the context of multi-label problems however, SSL is still under-explored. In this work we adapt Fix-Match to the multi-label scenario, specifically focusing on CheXpert, a multi-label chest X-ray classification dataset which is imbalanced and only partially labeled. Leveraging distribution alignment, our proposed method, ML-FixMatch+DA, achieves solid performance gains in SSL tasks (AUC: +2.6%) and in a missing label scenario (AUC: +1.9%). In contrast to previous work we achieve a performance gain on CheXpert using FixMatch. We show that in contrast to multiclass FixMatch, where distribution alignment is optional, it is essential for multi-label FixMatch to handle class imbalance and generate reliable (positive and negative) pseudo-labels. Our pseudo-label selection is based on a single threshold for all classes and handles imbalance with no prior knowledge on label distributions. Our adaptation keeps the simplicity of the original multi-class FixMatch with no added hyperparameters (even for imbalanced data) and demonstrates the feasibility of simple SSL for multi-label problems, filling a crucial gap in the literature.

KW - semi supervised learning

KW - CheXpert

KW - multi-label classification

KW - Missing Labels

KW - X-ray classification

KW - Semi-supervised multi-label learning

KW - Semi-supervised learning

KW - Imbalanced semi-supervised learning

KW - Medical Image Analysis

KW - Multi-Label Classification

UR - http://www.scopus.com/inward/record.url?scp=85206483137&partnerID=8YFLogxK

U2 - 10.1109/CVPRW63382.2024.00235

DO - 10.1109/CVPRW63382.2024.00235

M3 - Conference contribution

SN - 979-8-3503-6548-1

T3 - IEEE Computer Society Conference on Computer Vision and Pattern Recognition workshops

SP - 2295

EP - 2304

BT - Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

Von denselben Autoren