Leveraging Partial Labels for Cervical Lesion Classification via a Multilabel Approach
Early detection and diagnosis of cervical lesions is crucial for mitigating cervical cancer. Because of limited access to adequate annotations, existing studies on cervical lesion classification have primarily focused on generalized classification. This study introduces a multilabel approach to cerv...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10843686/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832584006106873856 |
|---|---|
| author | Margaret Dy Manalo Kota Aoki Shuqiong Wu Mariko Shindo Yutaka Ueda Yasushi Yagi |
| author_facet | Margaret Dy Manalo Kota Aoki Shuqiong Wu Mariko Shindo Yutaka Ueda Yasushi Yagi |
| author_sort | Margaret Dy Manalo |
| collection | DOAJ |
| description | Early detection and diagnosis of cervical lesions is crucial for mitigating cervical cancer. Because of limited access to adequate annotations, existing studies on cervical lesion classification have primarily focused on generalized classification. This study introduces a multilabel approach to cervical lesion classification through a data-centric framework. We first enhanced data quality via a comprehensive preprocessing pipeline that reduces the dataset to usable cervigrams. This refined dataset is then fed to a Vision Transformer, which performs per-class feature extraction and incorporates a part selection module that emphasizes critical areas within the cervigram, while integrating inter-class information within both early and late stages of the model. These procedures are conducted under the assumption of potential partial labeling. Upon evaluation, our proposed method outperformed existing models, achieving the highest ROC-AUC scores across all lesion grades. These findings suggest that the improved attention mechanisms led to enhanced localization of lesions, enabling focus on the fine details of each lesion. Overall, this study explores the potential of multilabel classification for advancing cervical lesion detection. |
| format | Article |
| id | doaj-art-edb830a2663f415ab58941081cae7e34 |
| institution | Kabale University |
| issn | 2169-3536 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj-art-edb830a2663f415ab58941081cae7e342025-01-28T00:01:20ZengIEEEIEEE Access2169-35362025-01-0113151881520010.1109/ACCESS.2025.353050410843686Leveraging Partial Labels for Cervical Lesion Classification via a Multilabel ApproachMargaret Dy Manalo0https://orcid.org/0000-0002-9447-8536Kota Aoki1https://orcid.org/0000-0002-3670-3109Shuqiong Wu2https://orcid.org/0000-0003-1501-9719Mariko Shindo3https://orcid.org/0000-0002-2864-0769Yutaka Ueda4Yasushi Yagi5https://orcid.org/0000-0002-3546-8071Department of Intelligent Media, SANKEN (Institute of Scientific and Industrial Research), Osaka University, Ibaraki, Osaka, JapanDepartment of Electrical Engineering and Computer Science, Faculty of Engineering, Tottori University, Tottori, Tottori, JapanDepartment of Intelligent Media, SANKEN (Institute of Scientific and Industrial Research), Osaka University, Ibaraki, Osaka, JapanDepartment of Obstetrics and Gynecology, Graduate School of Medicine, Osaka University, Suita, Osaka, JapanDepartment of Obstetrics and Gynecology, Graduate School of Medicine, Osaka University, Suita, Osaka, JapanDepartment of Intelligent Media, SANKEN (Institute of Scientific and Industrial Research), Osaka University, Ibaraki, Osaka, JapanEarly detection and diagnosis of cervical lesions is crucial for mitigating cervical cancer. Because of limited access to adequate annotations, existing studies on cervical lesion classification have primarily focused on generalized classification. This study introduces a multilabel approach to cervical lesion classification through a data-centric framework. We first enhanced data quality via a comprehensive preprocessing pipeline that reduces the dataset to usable cervigrams. This refined dataset is then fed to a Vision Transformer, which performs per-class feature extraction and incorporates a part selection module that emphasizes critical areas within the cervigram, while integrating inter-class information within both early and late stages of the model. These procedures are conducted under the assumption of potential partial labeling. Upon evaluation, our proposed method outperformed existing models, achieving the highest ROC-AUC scores across all lesion grades. These findings suggest that the improved attention mechanisms led to enhanced localization of lesions, enabling focus on the fine details of each lesion. Overall, this study explores the potential of multilabel classification for advancing cervical lesion detection.https://ieeexplore.ieee.org/document/10843686/Multilabel classificationpartial labelscervical cancercomputer-aided diagnosis |
| spellingShingle | Margaret Dy Manalo Kota Aoki Shuqiong Wu Mariko Shindo Yutaka Ueda Yasushi Yagi Leveraging Partial Labels for Cervical Lesion Classification via a Multilabel Approach IEEE Access Multilabel classification partial labels cervical cancer computer-aided diagnosis |
| title | Leveraging Partial Labels for Cervical Lesion Classification via a Multilabel Approach |
| title_full | Leveraging Partial Labels for Cervical Lesion Classification via a Multilabel Approach |
| title_fullStr | Leveraging Partial Labels for Cervical Lesion Classification via a Multilabel Approach |
| title_full_unstemmed | Leveraging Partial Labels for Cervical Lesion Classification via a Multilabel Approach |
| title_short | Leveraging Partial Labels for Cervical Lesion Classification via a Multilabel Approach |
| title_sort | leveraging partial labels for cervical lesion classification via a multilabel approach |
| topic | Multilabel classification partial labels cervical cancer computer-aided diagnosis |
| url | https://ieeexplore.ieee.org/document/10843686/ |
| work_keys_str_mv | AT margaretdymanalo leveragingpartiallabelsforcervicallesionclassificationviaamultilabelapproach AT kotaaoki leveragingpartiallabelsforcervicallesionclassificationviaamultilabelapproach AT shuqiongwu leveragingpartiallabelsforcervicallesionclassificationviaamultilabelapproach AT marikoshindo leveragingpartiallabelsforcervicallesionclassificationviaamultilabelapproach AT yutakaueda leveragingpartiallabelsforcervicallesionclassificationviaamultilabelapproach AT yasushiyagi leveragingpartiallabelsforcervicallesionclassificationviaamultilabelapproach |