Ordinal losses for classification of cervical cancer risk
Cervical cancer is the fourth leading cause of cancer-related deaths in women, especially in low to middle-income countries. Despite the outburst of recent scientific advances, there is no totally effective treatment, especially when diagnosed in an advanced stage. Screening tests, such as cytology...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
PeerJ Inc.
2021-04-01
|
| Series: | PeerJ Computer Science |
| Subjects: | |
| Online Access: | https://peerj.com/articles/cs-457.pdf |
| _version_ | 1818676270577221632 |
|---|---|
| author | Tomé Albuquerque Ricardo Cruz Jaime S. Cardoso |
| author_facet | Tomé Albuquerque Ricardo Cruz Jaime S. Cardoso |
| author_sort | Tomé Albuquerque |
| collection | DOAJ |
| description | Cervical cancer is the fourth leading cause of cancer-related deaths in women, especially in low to middle-income countries. Despite the outburst of recent scientific advances, there is no totally effective treatment, especially when diagnosed in an advanced stage. Screening tests, such as cytology or colposcopy, have been responsible for a substantial decrease in cervical cancer deaths. Cervical cancer automatic screening via Pap smear is a highly valuable cell imaging-based detection tool, where cells must be classified as being within one of a multitude of ordinal classes, ranging from abnormal to normal. Current approaches to ordinal inference for neural networks are found to not sufficiently take advantage of the ordinal problem or to be too uncompromising. A non-parametric ordinal loss for neuronal networks is proposed that promotes the output probabilities to follow a unimodal distribution. This is done by imposing a set of different constraints over all pairs of consecutive labels which allows for a more flexible decision boundary relative to approaches from the literature. Our proposed loss is contrasted against other methods from the literature by using a plethora of deep architectures. A first conclusion is the benefit of using non-parametric ordinal losses against parametric losses in cervical cancer risk prediction. Additionally, the proposed loss is found to be the top-performer in several cases. The best performing model scores an accuracy of 75.6% for seven classes and 81.3% for four classes. |
| first_indexed | 2024-12-17T08:40:49Z |
| format | Article |
| id | doaj.art-3e679c03ad3c489ca09469612b9aeca5 |
| institution | Directory Open Access Journal |
| issn | 2376-5992 |
| language | English |
| last_indexed | 2024-12-17T08:40:49Z |
| publishDate | 2021-04-01 |
| publisher | PeerJ Inc. |
| record_format | Article |
| series | PeerJ Computer Science |
| spelling | doaj.art-3e679c03ad3c489ca09469612b9aeca52022-12-21T21:56:21ZengPeerJ Inc.PeerJ Computer Science2376-59922021-04-017e45710.7717/peerj-cs.457Ordinal losses for classification of cervical cancer riskTomé Albuquerque0Ricardo Cruz1Jaime S. Cardoso2Institute for Systems and Computer Engineering, Technology and Science, Porto, PortugalInstitute for Systems and Computer Engineering, Technology and Science, Porto, PortugalInstitute for Systems and Computer Engineering, Technology and Science, Porto, PortugalCervical cancer is the fourth leading cause of cancer-related deaths in women, especially in low to middle-income countries. Despite the outburst of recent scientific advances, there is no totally effective treatment, especially when diagnosed in an advanced stage. Screening tests, such as cytology or colposcopy, have been responsible for a substantial decrease in cervical cancer deaths. Cervical cancer automatic screening via Pap smear is a highly valuable cell imaging-based detection tool, where cells must be classified as being within one of a multitude of ordinal classes, ranging from abnormal to normal. Current approaches to ordinal inference for neural networks are found to not sufficiently take advantage of the ordinal problem or to be too uncompromising. A non-parametric ordinal loss for neuronal networks is proposed that promotes the output probabilities to follow a unimodal distribution. This is done by imposing a set of different constraints over all pairs of consecutive labels which allows for a more flexible decision boundary relative to approaches from the literature. Our proposed loss is contrasted against other methods from the literature by using a plethora of deep architectures. A first conclusion is the benefit of using non-parametric ordinal losses against parametric losses in cervical cancer risk prediction. Additionally, the proposed loss is found to be the top-performer in several cases. The best performing model scores an accuracy of 75.6% for seven classes and 81.3% for four classes.https://peerj.com/articles/cs-457.pdfCervical cytologyConvolutional Neural networksDeep learningOrdinal classificationPap smear |
| spellingShingle | Tomé Albuquerque Ricardo Cruz Jaime S. Cardoso Ordinal losses for classification of cervical cancer risk PeerJ Computer Science Cervical cytology Convolutional Neural networks Deep learning Ordinal classification Pap smear |
| title | Ordinal losses for classification of cervical cancer risk |
| title_full | Ordinal losses for classification of cervical cancer risk |
| title_fullStr | Ordinal losses for classification of cervical cancer risk |
| title_full_unstemmed | Ordinal losses for classification of cervical cancer risk |
| title_short | Ordinal losses for classification of cervical cancer risk |
| title_sort | ordinal losses for classification of cervical cancer risk |
| topic | Cervical cytology Convolutional Neural networks Deep learning Ordinal classification Pap smear |
| url | https://peerj.com/articles/cs-457.pdf |
| work_keys_str_mv | AT tomealbuquerque ordinallossesforclassificationofcervicalcancerrisk AT ricardocruz ordinallossesforclassificationofcervicalcancerrisk AT jaimescardoso ordinallossesforclassificationofcervicalcancerrisk |