UKSSL: Underlying Knowledge Based Semi-Supervised Learning for Medical Image Classification
<italic>Goal:</italic> Deep learning techniques have made significant progress in medical image analysis. However, obtaining ground truth labels for unlabeled medical images is challenging as they often outnumber labeled images. Thus, training a high-performance model with limited labele...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2024-01-01
|
| Series: | IEEE Open Journal of Engineering in Medicine and Biology |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10218990/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832582316620251136 |
|---|---|
| author | Zeyu Ren Xiangyu Kong Yudong Zhang Shuihua Wang |
| author_facet | Zeyu Ren Xiangyu Kong Yudong Zhang Shuihua Wang |
| author_sort | Zeyu Ren |
| collection | DOAJ |
| description | <italic>Goal:</italic> Deep learning techniques have made significant progress in medical image analysis. However, obtaining ground truth labels for unlabeled medical images is challenging as they often outnumber labeled images. Thus, training a high-performance model with limited labeled data has become a crucial challenge. <italic>Methods:</italic> This study introduces an underlying knowledge-based semi-supervised framework called UKSSL, consisting of two components: MedCLR extracts feature representations from the unlabeled dataset; UKMLP utilizes the representation and fine-tunes it with the limited labeled dataset to classify the medical images. <italic>Results:</italic> UKSSL evaluates on the LC25000 and BCCD datasets, using only 50% labeled data. It gets precision, recall, F1-score, and accuracy of 98.9% on LC25000 and 94.3%, 94.5%, 94.3%, and 94.1% on BCCD, respectively. These results outperform other supervised-learning methods using 100% labeled data. <italic>Conclusions:</italic> The UKSSL can efficiently extract underlying knowledge from the unlabeled dataset and perform better using limited labeled medical images. |
| format | Article |
| id | doaj-art-85f6a349f1b4466da4c967a3ec49620b |
| institution | Kabale University |
| issn | 2644-1276 |
| language | English |
| publishDate | 2024-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Open Journal of Engineering in Medicine and Biology |
| spelling | doaj-art-85f6a349f1b4466da4c967a3ec49620b2025-01-30T00:03:48ZengIEEEIEEE Open Journal of Engineering in Medicine and Biology2644-12762024-01-01545946610.1109/OJEMB.2023.330519010218990UKSSL: Underlying Knowledge Based Semi-Supervised Learning for Medical Image ClassificationZeyu Ren0https://orcid.org/0000-0003-2303-5663Xiangyu Kong1https://orcid.org/0009-0001-5365-9445Yudong Zhang2https://orcid.org/0000-0002-4870-1493Shuihua Wang3https://orcid.org/0000-0003-2238-6808University of Leicester, Leicester, U.K.University of Leicester, Leicester, U.K.University of Leicester, Leicester, U.K.University of Leicester, Leicester, U.K.<italic>Goal:</italic> Deep learning techniques have made significant progress in medical image analysis. However, obtaining ground truth labels for unlabeled medical images is challenging as they often outnumber labeled images. Thus, training a high-performance model with limited labeled data has become a crucial challenge. <italic>Methods:</italic> This study introduces an underlying knowledge-based semi-supervised framework called UKSSL, consisting of two components: MedCLR extracts feature representations from the unlabeled dataset; UKMLP utilizes the representation and fine-tunes it with the limited labeled dataset to classify the medical images. <italic>Results:</italic> UKSSL evaluates on the LC25000 and BCCD datasets, using only 50% labeled data. It gets precision, recall, F1-score, and accuracy of 98.9% on LC25000 and 94.3%, 94.5%, 94.3%, and 94.1% on BCCD, respectively. These results outperform other supervised-learning methods using 100% labeled data. <italic>Conclusions:</italic> The UKSSL can efficiently extract underlying knowledge from the unlabeled dataset and perform better using limited labeled medical images.https://ieeexplore.ieee.org/document/10218990/Deep learningself-supervised learningmedical image analysissemi-supervised learningimage classification |
| spellingShingle | Zeyu Ren Xiangyu Kong Yudong Zhang Shuihua Wang UKSSL: Underlying Knowledge Based Semi-Supervised Learning for Medical Image Classification IEEE Open Journal of Engineering in Medicine and Biology Deep learning self-supervised learning medical image analysis semi-supervised learning image classification |
| title | UKSSL: Underlying Knowledge Based Semi-Supervised Learning for Medical Image Classification |
| title_full | UKSSL: Underlying Knowledge Based Semi-Supervised Learning for Medical Image Classification |
| title_fullStr | UKSSL: Underlying Knowledge Based Semi-Supervised Learning for Medical Image Classification |
| title_full_unstemmed | UKSSL: Underlying Knowledge Based Semi-Supervised Learning for Medical Image Classification |
| title_short | UKSSL: Underlying Knowledge Based Semi-Supervised Learning for Medical Image Classification |
| title_sort | ukssl underlying knowledge based semi supervised learning for medical image classification |
| topic | Deep learning self-supervised learning medical image analysis semi-supervised learning image classification |
| url | https://ieeexplore.ieee.org/document/10218990/ |
| work_keys_str_mv | AT zeyuren uksslunderlyingknowledgebasedsemisupervisedlearningformedicalimageclassification AT xiangyukong uksslunderlyingknowledgebasedsemisupervisedlearningformedicalimageclassification AT yudongzhang uksslunderlyingknowledgebasedsemisupervisedlearningformedicalimageclassification AT shuihuawang uksslunderlyingknowledgebasedsemisupervisedlearningformedicalimageclassification |