Ensemble Learning-Based Non-Invasive Anemia Detection Using Conjunctival Image Analysis

Anemia is a global health issue characterized by low hemoglobin levels, leading to reduced oxygen transport in the body. Traditional diagnosis relies on invasive blood tests, which can be costly and inaccessible. This study proposes a machine learning-based ensemble framework for non-invasive anemia detection using conjunctival images. Various classifiers, including SVM, Random Forest (RF), XGBoost, and ANN, were trained and evaluated. Ensemble learning techniques such as bagging, boosting, stacking, and voting were applied, with the stacking ensemble (RF + ANN + XGB) achieving the highest accuracy (95.77%). Extra Trees emerged as the best-performing individual model with 96.48% accuracy. The results confirm that ensemble models significantly improve classification performance over individual classifiers. Future work should integrate multi-modal approaches to further enhance diagnostic accuracy. This study highlights the potential of ensemble learning for scalable, non-invasive anemia detection in clinical settings.

Anemia detection, machine learning, ensemble learning, stacking, conjunctival images, Random Forest, XGBoost, Extra Trees, medical image classification

1. Kassebaum, N. J., Jasrasaria, R., Naghavi, M., et al. (2014). A systematic analysis of global anemia burden from 1990 to 2010. Blood, 123(5), 615-624. https://doi.org/10.1182/blood-2013-06-508325 [Google Scholar] [Crossref]

2. McLean, E., Cogswell, M., Egli, I., Wojdyla, D., & de Benoist, B. (2009). Worldwide prevalence of anemia, WHO Vitamin and Mineral Nutrition Information System, 1993–2005. Public Health Nutrition, 12(4), 444-454. https://doi.org/10.1017/S1368980008002401 [Google Scholar] [Crossref]

3. von Schenck, H., Falkensson, M., & Lundberg, B. (1986). Evaluation of “HemoCue,” a new device for determining hemoglobin. Clinical Chemistry, 32(3), 526-529. [Google Scholar] [Crossref]

4. https://doi.org/10.1093/clinchem/32.3.526 [Google Scholar] [Crossref]

5. Asare, J. W., Appiahene, P., Donkoh, E. T., & Dimauro, G. (2023). Detection of anemia using conjunctiva images: A smartphone application approach. Informatics in Medicine Unlocked, 40, 101283. https://doi.org/10.1016/j.imu.2023.101283 [Google Scholar] [Crossref]

6. Asare, J., Appiahene, P., Donkoh, E., & Dimauro, G. (2023). Iron deficiency anemia detection using machine learning models: A comparative study of fingernails, palm, and conjunctiva of the eye images. Engineering Reports, 5(11), e12667. https://doi.org/10.1002/eng2.12667 [Google Scholar] [Crossref]

7. Appiahene, P., Dogbe, S. S. D., Kobina, E. E. Y., Dartey, P. S., Afrifa, S., Donkoh, E. T., & Asare, J. W. (2023). Application of ensemble models approach in anemia detection using images of the palpable palm. Medicine in Novel Technology and Devices, 20, 100269. [Google Scholar] [Crossref]

8. https://doi.org/10.1016/j.medntd.2023.100269 [Google Scholar] [Crossref]

9. GBD 2021 Anaemia Collaborators. (2023). Prevalence, years lived with disability, and trends in anaemia burden by severity and cause, 1990–2021: findings from the Global Burden of Disease Study 2021. The Lancet Haematology, 10(8), e574-e586. https://doi.org/10.1016/S2352-3026(23)00160-6 [Google Scholar] [Crossref]

10. Asare, Justice Williams; Appiahene, Peter; Donkoh, Emmanuel. (2023). CP-AnemiC (A Conjunctival Pallor) Dataset from Ghana. Mendeley Data, V1. https://doi.org/10.17632/m53vz6b7fx.1 [Google Scholar] [Crossref]

11. Collings, S., Ellingson, K. D., Keeler, C., et al. (2016). Non-invasive detection of anaemia using digital photographs of the conjunctiva. PLoS One, 11(4), e0153286. [Google Scholar] [Crossref]

12. https://doi.org/10.1371/journal.pone.0153286 [Google Scholar] [Crossref]

13. Appiahene, P., Chaturvedi, K., Asare, J. W., Donkoh, E. T., & Prasad, M. (2023). CP-AnemiC: A conjunctival pallor dataset and benchmark for anemia detection in children. Medicine in Novel Technology and Devices, 18, 100244. https://doi.org/10.1016/j.medntd.2023.100244 [Google Scholar] [Crossref]

14. Dimauro, G., Bevilacqua, V., Colizzi, L., et al. (2021). A machine learning approach for anemia detection based on conjunctiva images. Computer Methods and Programs in Biomedicine, 200, 105911. https://doi.org/10.1016/j.cmpb.2020.105911 [Google Scholar] [Crossref]

15. Mannino, R. G., Myers, D. R., Tyburski, E. A., et al. (2018). Smartphone app for non-invasive detection of anemia using only patient-sourced photos. Nature Communications, 9(1), 4924. [Google Scholar] [Crossref]

16. https://doi.org/10.1038/s41467-018-07262-2 [Google Scholar] [Crossref]

17. Shao, J., Wong, J. Y., Poljak, M., et al. (2017). HemaApp: Noninvasive blood screening of hemoglobin using smartphone cameras. Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies, 1(3), 97. https://doi.org/10.1145/3130952 [Google Scholar] [Crossref]

18. Muljono, S. A., Wulandari, H. A., Azies, M., Naufal, W. A., Prasetyanto, F. A., & Zahra, F. A. (2024). Breaking boundaries in diagnosis: Non-invasive anemia detection empowered by AI. IEEE Access, 12, 9292-9307. https://doi.org/10.1109/ACCESS.2024.3353788 [Google Scholar] [Crossref]

19. Asare, Justice Williams; Appiahene, Peter; Donkoh, Emmanuel. (2023). CP-AnemiC (A Conjunctival Pallor) Dataset from Ghana. Mendeley Data, V1. https://doi.org/10.17632/m53vz6b7fx.1 [Google Scholar] [Crossref]

20. Ali, M. A., Dornaika, F., Cherry, A., et al. (2024). Advancing medical image processing: A comparative analysis of ‘Naturalize’ and traditional augmentation techniques for enhanced diagnostic accuracy. IC2SPM, 138–143. https://doi.org/10.1109/ic2spm62723.2024.10841335 [Google Scholar] [Crossref]

• Classifying News Articles on Nigerian Newspapers Using Machine Learning
• Approaching Early Dyslexia Prediction with Ensemble Machine Learning: A Secondary Empirical Analysis of the Rello Et Al. Public Behavioural Dataset
• Machine Learning Modelling of Geostationary SBAS Satellite Orbit Errors Observed from an Equatorial West African Station
• A Comparative Study of CNN and Vision Transformer Models in Alzheimer's Detection
• A Stacking Ensemble Machine Learning Framework for Predicting Academic Performance and Identifying At-Risk Students in Rural Secondary Schools: Evidence from Aurangabad District, Maharashtra