Mineral classification using Dense-Net architectures: leveraging deep feature extraction for enhanced accuracy

Year 2025, Volume: 9 Issue: 3, 201 - 210, 25.12.2025

Youcef Attallah , Ehlem Zigh , Seydi Kaçmaz , Amine Maazouzi

https://doi.org/10.35860/iarej.1687724

Abstract

Mineral classification is a fundamental task in geoscience, where deep learning models have shown remarkable potential in automating and improving identification accuracy. In this study, we conduct a comprehensive comparative analysis of three DenseNet architectures—DenseNet-121, DenseNet-169, and DenseNet-201—applied to the Minet v2 dataset, which contains 5,640 images spanning seven mineral classes. Unlike prior studies limited to single models or earlier datasets, our work provides a systematic evaluation of performance, model complexity, and computational efficiency. The experimental results demonstrate that DenseNet-201 achieves the highest overall accuracy (95.76 %), precision (94.16 %), recall (94.84 %), and the lowest error rate (4.24 %), while DenseNet-169 provides a strong balance between accuracy (94.70 %) and efficiency. DenseNet-121, although lighter with 8.1 million parameters, achieves a lower accuracy of 92.23 %. These findings confirm the advantage of deeper DenseNet variants for fine-grained mineral recognition, while also highlighting trade-offs relevant to resource-constrained applications. By presenting the first detailed comparative analysis of DenseNet architectures on the Minet v2 benchmark, this study contributes valuable insights for the development of robust, scalable, and efficient mineral classification systems.

Keywords

Deep Learning , DenseNet , Mineral Classifiation , Image Analysis , Transfer Learning

Thanks

The authors would like to express their sincere gratitude to the Department of Electrical and Electronics Engineering at Gaziantep University, Türkiye, for supporting this research within the framework of the Erasmus+ Program. This support has been essential to the successful completion of this work.

References

• 1. Qingyu, S. and Mukherjee, Applied mineralogy: applications in industry and environment. Springer Science & Business Media, 2012.
• 2. Egger, A.E., Teaching the process of science. in Proceedings of the 2009 Process of Science Workshop. 2009. Science Education Resource Centre, Carleton College.
• 3. Huang, G., Liu, Z., Van Der Maaten, L., et al., Densely connected convolutional networks. in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2017. p. 4700–4708.
• 4. Abdurahman, F., Handwritten Amharic character recognition system using convolutional neural networks. Engineering Sciences, 2019. 14(2): p. 71–87.
• 5. Güçlü, S., Özdemir, D., and Saraoğlu, H.M., Une nouvelle méthode de détection d’anomalies basée sur DenseNet utilisant des images radiographiques du poignet et de l’avant-bras. ESTUDAM Bilişim, 2024. 5(2): p. 18–29.
• 6. Soui, M. and Haddad, Z., Deep learning-based model using DenseNet201 for mobile user interface evaluation. International Journal of Human Computer Interaction, 2023. 39(9): p. 1981–1994.
• 7. Thompson, S., Fueten, F., and Bockus, D., Mineral identification using artificial neural networks and the rotating polarizer stage. Computers & Geosciences, 2001. 27(9): p. 1081–1089.
• 8. Zhang, Y., Li, M., Han, S., et al., Intelligent identification for rock mineral microscopic images using ensemble machine learning algorithms. Sensors, 2019. 19(18): 3914.
• 9. Suhasini, B.R.C., Minerals classification using convolutional neural network. International Research Journal of Engineering and Technology (IRJET), 2021. 8(2): p. 1686–1690.
• 10. Brempong, E.A., Agangiba, M., and Aikins, D., Minet: A convolutional neural network for identifying and categorising minerals. arXiv preprint arXiv:2111.11260, 2021.
• 11. Wang, S.-H. and Zhang, Y.-D., DenseNet-201-based deep neural network with composite learning factor and precomputation for multiple sclerosis classification. ACM Transactions on Multimedia Computing, Communications, and Applications (TOMM), 2020. 16(2s): p. 1–19.
• 12. Shorten, C. and Khoshgoftaar, T.M., A survey on image data augmentation for deep learning. Journal of Big Data, 2019. 6(1): p. 1–48.
• 13. Torunoglu Selamet, D. and Cebiroğlu Eryiğit, G., Désambiguïsation contextuelle semi-supervisée pour l’augmentation des données. TBV-BBMD, 2021. 14(1): p. 34–46.
• 14. Günay, M. and Köseoğlu, M., Detection of circuit components on hand-drawn circuit images by using faster R-CNN method. International Advanced Research Engineering Journal (IAREJ), 2021. 5(3): p. 372–378.
• 15. Khubchandani, V., Image caption generator using DenseNet201 and ResNet50. International Journal of Future Computer and Communication, 2024. 13(3).
• 16. Attallah, Y., Minerals identification classification Minet v2 (large dataset — 7 classes). 2022. Available from: https://www.kaggle.com/datasets/youcefattallah97/minerals-identification-classification
• 17. Alp, S., Transfer learning for Turkish cuisine classification. BSJ Engineering Sciences, 2024. 7(6): p. 1302–1309.
• 18. He, K., Zhang, X., Ren, S., et al., Deep residual learning for image recognition. in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2016. p. 770–778.
• 19. Bal, F. and Kayaalp, F., Review of machine learning and deep learning models in agriculture. International Advanced Research Engineering Journal (IAREJ), 2021. 5(2): p. 309–323.
• 20. Kingma, D.P. and Ba, J., Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980, 2014.
• 21. Sevli, O. and Kemaloğlu, N., Turkish sign language digits classification with CNN using different optimizers. International Advanced Research Engineering Journal (IAREJ), 2020. 4(3): p. 200–207.
• 22. Goodfellow, I., Bengio, Y., and Courville, A., Deep learning. MIT Press, 2016.
• 23. Kulkarni, A., Chong, D., and Batarseh, F.A., Foundations of data imbalance and solutions for a data democracy. in Data Democracy. Elsevier, 2020. p. 83–106.
• 24. Micikevicius, P., Narang, S., Alben, J., et al., Mixed precision training. in International Conference on Learning Representations (ICLR). 2018.
• 25. Liu, Z., Mao, H., Wu, C.-Y., Feichtenhofer, C., Darrell, T., and Xie, S., A convnet for the 2020s. in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2022. p. 11976–11986.
• 26. Tan, M. and Le, Q., EfficientNetV2: Smaller models and faster training. in Proceedings of the International Conference on Machine Learning (ICML). 2021. p. 10096–10106. PMLR.
• 27. Altun Güven, S. and Toptaş, B., Using up-to-date GAN methods for aerial images. DUJE, 2024. 15(1): p. 87–97.