A Systematic Review on the Impact Relationship of Data Quality on Algorithmic Justice for Image Classification

• Maristela de Freitas Riquelme UFPB
• Lucas Freire de Lima UFPB
• Luiz Fernando F. P. de Lima CESAR
• Danielle Rousy Dias Ricarte UFPB

DOI: https://doi.org/10.5753/sbcas_estendido.2024.2770

Abstract

As medical image classification systems become more widespread, the debate surrounding their fairness and impartiality intensifies. Seeking to understand how this issue is being discussed, a systematic review was conducted on the impact of data quality on biases in machine learning systems for medical image classification. After analyzing the articles, methods were identified to ensure the quality of the datasets. It is concluded that the quality of the dataset impacts the performance of the models, potentially leading to incorrect or imprecise clinical diagnoses.

References

Arora, A. et al. (2023) “The value of standards for health datasets in artificial intelligence-based applications”, Em: Nature Medicine, 29, 2929-2938, DOI: 10.1038/s41591-023-02608-w

Band, S. S. et al. (2023) “Application of explainable artificial intelligence in medical health: A systematic review of interpretability methods”, Em: Informatics in Medicine Unlocked, 40, 101286, DOI: 10.1016/j.imu.2023.101286

Dash, S., Vineeth, N. B. and Sharma, A. (2022) “Evaluating and Mitigating Bias in Image Classifiers: A Causal Perspective Using Counterfactuals”, Em: 2022 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), Waikoloa, HI, USA, p. 3879-3888, DOI: 10.1109/WACV51458.2022.00393.

El-Sappagh, S. et al. (2023) “Trustworthy artificial intelligence in Alzheimer’s disease: state of the art, opportunities, and challenges”, Em: Artificial Intelligence Review, 56, p. 11149 – 11296, DOI: 10.1007/s10462-023-10415-5

Kitchenham, B. (2004) “Procedures for Performing Systematic Reviews”, [link]

Lei, J. et al. (2023) “Category-aware feature attribution for Self-Optimizing medical image classification”, Em: Displays, 77, 102397, DOI: 10.1016/j.displa.2023.102397

Pandl, K. D. et al. (2021) “Trustworthy machine learning for health care: scalable data valuation with the shapley value”, Em: CHIL '21: Proceedings of the Conference on Health, Inference, and Learning, p. 47 - 57, DOI: 10.1145/3450439.3451861

Tian, F. et al. (2022) “Face Recognition Fairness Assessment based on Data Augmentation: An Empirical Study”, Em: 2022 IEEE 22nd International Conference on Software Quality, Reliability, and Security Companion (QRS-C), Guangzhou, China, p. 315-318, DOI: 10.1109/QRS-C57518.2022.00053.

Yang, J. et al. (2023) “Algorithmic fairness and bias mitigation for clinical machine learning with deep reinforcement learning”, Em: Nature Machine Intelligence, 5, 884-894, DOI: 10.1038/s42256-023-00697-3.