tomealbuquerque / ordinal-losses

Work involving multiple losses for ordinal classification

Ordinal Losses for Classification of Cervical Cancer Risk

https://peerj.com/articles/cs-457/

by Tomé Albuquerque, Ricardo Cruz, Jaime S. Cardoso

Introduction

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 7 classes and 81.3% for 4 classes.

Usage

1. Run preprocess.py to generate the folds.
2. Run train.py to train the models you want.
3. Run evaluate.py to generate results table.

Code organization

• data: avaiable at: http://mde-lab.aegean.gr/index.php/downloads
• train.py: train the different models with the different ordinal losses and outputs probabilities.
• evaluate.py: generate latex tables with results using the output probabilities.

Citation

If you find this work useful for your research, please cite our paper:

@article{albuquerque2021ordinal,
 title = {Ordinal losses for classification of cervical cancer risk},
 author = {Albuquerque, Tomé and Cruz, Ricardo and Cardoso, Jaime S.},
 year = 2021,
 month = 4,
 keywords = {Cervical cytology, Convolutional Neural networks, Deep learning, Ordinal classification, Pap smear},
 volume = 7,
 pages = {e457},
 journal = {PeerJ Computer Science},
 issn = {2376-5992},
 url = {https://doi.org/10.7717/peerj-cs.457},
 doi = {10.7717/peerj-cs.457}
}

If you have any questions about our work, please do not hesitate to contact [e-mail]