Analysis of Multi-type Barcode Images Using Hybrid CNN-Transformer Model for Water Quality Classification

Abstract

Water quality is a critical factor for both human health and aquatic ecosystems. An inaccurate assessment of water quality can have serious consequences, including risks to drinking water supplies and adverse effects on aquatic life. Therefore, there is growing demand for rapid, reliable methods to estimate water quality. In recent years, the application of deep learning (DL)-based methods for assessing water quality has increased significantly. DL methods achieve high classification performance by autonomously learning relationships within complex water quality datasets. This study aims to classify two-dimensional barcode-based visual representations of water quality using advanced DL architectures. The study used a tabular water-quality dataset comprising physicochemical parameters, which were subsequently transformed into QR, Aztec, and Data Matrix code images. These barcode datasets were initially evaluated using transfer learning with the MobileNet, MobileNetV2, and MobileNetV3Small Convolutional Neural Network (CNN) architectures, with a 5-fold cross-validation approach. Subsequently, the MobileNetV3Small model, which achieved the highest classification performance, served as the backbone of the CNN-Transformer architecture, enabling image classification with this proposed model. Based on the experimental results, the highest average performance was achieved on the Aztec code dataset. Specifically, the MobileNetV3Small model attained an accuracy of 91.95±1.19%, whereas the MobileNetV3Small-Transformer model reached 97.72±2.19%. Additionally, the proposed MobileNetV3Small-Transformer model yielded accuracy rates of 92.65±5.88% on the QR code dataset and 90.44±7.76% on the Data Matrix code dataset. This approach presents an alternative methodology for evaluating tabular data through visual representations and DL techniques in the field of environmental data analytics.

Keywords

• Water quality
• tabular data
• barcode images
• convolutional neural networks
• transformer architecture

Su Kalitesinin Sınıflandırılması için Hibrit CNN-Transformer Modeli ile Çok Türlü Barkod Görüntülerinin Analizi

Abstract

Su kalitesi, insan sağlığı ve suda bulunan canlıların yaşamları için oldukça önemli bir faktördür. Su kalitesinin yanlış değerlendirilmesi, su temini ve su ekosistemleri bakımından olumsuz sonuçlar oluşturabilmektedir. Bu nedenle su kalitesinin doğru ve hızlı biçimde belirlenmesi önemli bir gerekliliktir. Son yıllarda, su kalitesinin değerlendirilmesi amacıyla derin öğrenme (DL) tabanlı yöntemlerin kullanımı giderek artmıştır. DL yöntemleri, karmaşık su kalitesi verileri arasındaki ilişkileri otomatik olarak öğrenebilme yetenekleri sayesinde, yüksek sınıflandırma performansı sunmaktadır. Bu bağlamda, çalışmanın amacı, su kalitesine ait iki boyutlu barkod tabanlı görsel temsillerin DL mimarileriyle sınıflandırılmasıdır. Çalışmada suya ait fizikokimyasal parametreler içeren bir tablosal su kalitesi veri seti kullanılmış ve veriler QR, Aztec ve Data Matrix kod görsellerine dönüştürülmüştür. Elde edilen barkod veri setleri, öncelikle aktarım öğrenme yaklaşımı kullanılarak MobileNet, MobileNetV2 ve MobileNetV3Small Evrişimsel Sinir Ağı (CNN) mimarileri ile 5-kat çapraz doğrulama yöntemi kapsamında değerlendirilmiştir. Ardından, en yüksek sınıflandırma performansı elde eden MobileNetV3Small modeli, CNN-Transformer mimarisinin omurgası olarak kullanılmış ve önerilen bu modelle görüntüler sınıflandırılmıştır. Deneysel sonuçlara göre, ortalama olarak en yüksek performans sonuçları Aztec kod veri setinde elde edilmiştir. Bu veri setinde MobileNetV3Small modeli %91,95±1,19 ve MobileNetV3Small-Transformer modeli %97,72±2,19 doğruluk değerleri elde etmiştir. Öte yandan, önerilen MobileNetV3Small-Transformer modeli QR kod veri setinde %92,65±5,88 ve Data Matrix kod veri setinde %90,44±7,76 doğruluk değerlerine ulaşmıştır. Önerilen yaklaşım, çevresel veri analitiği alanında tablosal verilerin görsel temsiller ve DL yöntemiyle değerlendirilmesine yönelik alternatif bir çözüm sunmaktadır.

Keywords

• Su kalitesi
• tablosal veriler
• barkod görüntüleri
• evrişimsel sinir ağları
• transformer mimarisi

References

1. Ahmed, A. N., Othman, F. B., Afan, H. A., Ibrahim, R. K., Fai, C. M., Hossain, M. S., Ehteram, M. & Elshafie, A. (2019a). Machine learning methods for better water quality prediction. Journal of Hydrology, 578, 124084.
2. Ahmed, U., Mumtaz, R., Anwar, H., Shah, A. A., Irfan, R. & García-Nieto, J. (2019b). Efficient water quality prediction using supervised machine learning. Water, 11 (11), 2210.
3. Ali, M., Yaseen, M., Ali, S. & Kim, H.-C. (2025). Deep Learning-Based Approach for Microscopic Algae Classification with Grad-CAM Interpretability. Electronics, 14 (3), 442.
4. Arabelli, R., Bernatin, T. & Veeramsetty, V. (2025). Water quality assessment for aquaculture using deep neural network. Desalination and Water Treatment, 101016.
5. Azrour, M., Mabrouki, J., Fattah, G., Guezzaz, A. & Aziz, F. (2022). Machine learning algorithms for efficient water quality prediction. Modeling Earth Systems and Environment, 8 (2), 2793–2801.
6. Bhatlawande, S., Mohole, S., Pahade, Y., Shilaskar, S. & Munjale, A. (2025). A Vision-Based System for Real-Time Monitoring of Water Quality in Natural and Artificial Reservoirs. 2025 3rd International Conference on Intelligent Data Communication Technologies and Internet of Things (IDCIoT), 1102–1107.
7. Chellaiah, C., Anbalagan, S., Swaminathan, D., Chowdhury, S., Kadhila, T., Shopati, A. K., Shangdiar, S., Sharma, B. & Amesho, K. T. (2024). Integrating deep learning techniques for effective river water quality monitoring and management. Journal of Environmental Management, 370, 122477.
8. de Fleury, M., Grippa, M., Brandt, M., Fensholt, R., Reiner, F., Kovacs, G. M. & Kergoat, L. (2025). Highly turbid and eutrophic small water bodies in West Africa well identified by a CNN U-Net algorithm. Remote Sensing Applications: Society and Environment, 37, 101412.

9. Dilmi, S. & Ladjal, M. (2021). A novel approach for water quality classification based on the integration of deep learning and feature extraction techniques. Chemometrics and Intelligent Laboratory Systems, 214, 104329.
10. Dosovitskiy, A. (2020). An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929.
11. Duan, G., Zhang, S., Shang, Y., Shao, Y. & Han, Y. (2025). Research on a Multi-Type Barcode Defect Detection Model Based on Machine Vision. Applied Sciences, 15 (15), 8176.
12. Elmotawakkil, A., Enneya, N., Bhagat, S. K., Ouda, M. M. & Kumar, V. (2025). Advanced machine learning models for robust prediction of water quality index and classification. Journal of Hydroinformatics, 27 (2), 299–319.
13. Erol Doğan, G. & Uzbaş, B. (2023). Diagnosis of COVID-19 from blood parameters using convolutional neural network. Soft Computing, 27 (15), 10555–10570.
14. Essamlali, I., Nhaila, H. & El Khaili, M. (2024). Advances in machine learning and IoT for water quality monitoring: A comprehensive review. Heliyon, 10 (6).
15. Fernández-Manteca, M. G., García, B. G., Álvarez, S. D., Gómez-Galdós, C., Pérez-Asensio, A., Algorri, J. F., Monteoliva, A. P., López-Higuera, J. M., Rodríguez-Cobo, L. & Ocampo-Sosa, A. A. (2025). Comprehensive Raman spectroscopy analysis for differentiating toxic cyanobacteria through multichannel 1D-CNNs and SHAP-based explainability. Talanta, 292, 127845.
16. Gómez-Martínez, V., Lara-Abelenda, F. J., Peiro-Corbacho, P., Chushig-Muzo, D., Granja, C. & Soguero-Ruiz, C. (2024). LM-IGTD: a 2D image generator for low-dimensional and mixed-type tabular data to leverage the potential of convolutional neural networks. arXiv preprint arXiv:2406.14566.
17. Gour, A. R., Kumar, R., Reddy, V., Amruthaluru, U. D., Hariharan, S. & Kukreja, V. (2024). Detection of Water Quality for Health Monitoring through CNN Image Analysis. 2024 4th International Conference on Pervasive Computing and Social Networking (ICPCSN), 31–36.
18. Gür, Y. E., Toğaçar, M. & Solak, B. (2025). Integration of CNN Models and Machine Learning Methods in Credit Score Classification: 2D Image Transformation and Feature Extraction. Computational Economics, 65, 2991-3035.
19. Haq, K. R. A. & Harigovindan, V. (2022). Water quality prediction for smart aquaculture using hybrid deep learning models. Ieee Access, 10, 60078–60098.
20. He, K., Zhang, X., Ren, S. & Sun, J. (2016). Deep residual learning for image recognition. Proceedings of the IEEE conference on computer vision and pattern recognition, 770–778.
21. Howard, A., Sandler, M., Chu, G., Chen, L.-C., Chen, B., Tan, M., Wang, W., Zhu, Y., Pang, R. & Vasudevan, V. (2019). Searching for mobilenetv3. Proceedings of the IEEE/CVF international conference on computer vision, 1314–1324.
22. Howard, A. G., Zhu, M., Chen, B., Kalenichenko, D., Wang, W., Weyand, T., Andreetto, M. & Adam, H. (2017). Mobilenets: Efficient convolutional neural networks for mobile vision applications. arXiv preprint arXiv:1704.04861.
23. Karpagam, V. & Christy, S. (2024). Deep learning-based water quality index classification using stacked ensemble variational mode decomposition. Environmental Research Communications, 6 (6), 065005.
24. Khadilkar, A. & Stamp, M. (2025). Image-Based Malware Classification Using QR and Aztec Codes. In: Machine Learning, Deep Learning and AI for Cybersecurity. Eds: Springer, p. 3–35.
25. Kirui, J. (2024). Machine learning models for drinking water quality classification. 2024 International Conference on Control, Automation and Diagnosis (ICCAD), 1–5.
26. König, T., Wagner, F., Liebschner, M. & Kley, M. (2025). Generative Adversarial Networks in Fault Diagnosis and an implementation approach: A review. IEEE Sensors Journal, 25 (14), 26289-26299.
27. Kumar, S., Pande, M. & Damle, A. Y. (2025). Comparative Analysis of Lion and AdamW Optimizers for Cross-Encoder Reranking with MiniLM, GTE, and ModernBERT. arXiv preprint arXiv:2506.18297.
28. Ladino, K. S. & Sama, M. P. (2025). Optimizing matrix barcode ground control points for automated location detection in UAS-based remote sensing. Computers and Electronics in Agriculture, 237, 110717.
29. Makumbura, R. K., Mampitiya, L., Rathnayake, N., Meddage, D., Henna, S., Dang, T. L., Hoshino, Y. & Rathnayake, U. (2024). Advancing water quality assessment and prediction using machine learning models, coupled with explainable artificial intelligence (XAI) techniques like shapley additive explanations (SHAP) for interpreting the black-box nature. Results in Engineering, 23, 102831.
30. Nie, Y., Chen, Y., Guo, J., Li, S., Xiao, Y., Gong, W. & Lan, R. (2025). An improved CNN model in image classification application on water turbidity. Scientific reports, 15 (1), 11264.
31. Noever, D. & McKee, F. (2025). Dueling QR Codes: The Hyding of Dr. Jeckyl. arXiv preprint arXiv:2503.13458.
32. Prasad, D. V. V., Venkataramana, L. Y., Kumar, P. S., Prasannamedha, G., Harshana, S., Srividya, S. J., Harrinei, K. & Indraganti, S. (2022). Analysis and prediction of water quality using deep learning and auto deep learning techniques. Science of the Total Environment, 821, 153311.
33. Ravindran, K. K. G., Della Monica, C., Atzori, G., Lambert, D., Hassanin, H., Revell, V. & Dijk, D.-J. (2023). Three contactless sleep technologies compared with actigraphy and polysomnography in a heterogeneous group of older men and women in a model of mild sleep disturbance: sleep laboratory study. JMIR mHealth and uHealth, 11, e46338.
34. Querini, M. & Italiano, G. F. (2012). Facial biometrics for 2D barcodes. 2012 Federated Conference on Computer Science and Information Systems (FedCSIS), 755–762.
35. Roy, B., Faruk, M. F., Islam, M. N., Srizon, A. Y., Hasan, S. M., Al Mamun, M., Hossain, M. R. & Hossain, M. F. (2024). A cutting-edge ensemble of Vision Transformer and ResNet101v2 based transfer learning for the precise classification of leukemia sub-types from peripheral blood smear images. 2024 6th International Conference on Electrical Engineering and Information & Communication Technology (ICEEICT), 49–54.
36. Sandler, M., Howard, A., Zhu, M., Zhmoginov, A. & Chen, L.-C. (2018). Mobilenetv2: Inverted residuals and linear bottlenecks. Proceedings of the IEEE conference on computer vision and pattern recognition, 4510–4520.
37. Sawant, P. V. & Patil, Y. M. (2024). Water Quality Monitoring Using Machine Learning Model. J. Electr. Syst, 20 (10s), 5686–5694.
38. Taha, B., Liza, F. R., Moniruzzaman, M. & Samsuzzaman, M. (2022). Comparative Analysis of Fine-Tuned MobileNet versions to Detect Brain Tumors from MRI images. 2022 4th International Conference on Sustainable Technologies for Industry 4.0 (STI), 1–6.
39. Tanwar, V., Sharma, B., Yadav, D. P. & Dwivedi, A. D. (2025). Enhancing blood cell diagnosis using hybrid residual and dual block transformer network. Bioengineering, 12 (2), 98.
40. Toğaçar, M. (2026). A Novel Approach to Malware Detection: Converting Hardware Memory Data to 2D Barcodes Using Mobile Networks. Journal of Network and Systems Management, 34 (1), 1–30.
41. Toumi, S., Lekmine, S., Touzout, N., Moussa, H., Elboughdiri, N., Boudraa, R., Benslama, O., Kebir, M., Danish, S. & Zhang, J. (2024). Harnessing deep learning for real-time water quality assessment: a sustainable solution. Water, 16 (23), 3380.
42. Turan, T. (2023). Makine öğrenmesi algoritmaları ile su kalitesi ve içilebilirlik tahmini. Uluborlu Mesleki Bilimler Dergisi, 6 (2), 65–80.
43. Veeramsetty, V., Arabelli, R. & T., B. (2024), Aquaculture - Water Quality Dataset, https://data.mendeley.com/datasets/y78ty2g293/1.
44. Wang, Z. & Oates, T. (2015a). Encoding time series as images for visual inspection and classification using tiled convolutional neural networks. Workshops at the twenty-ninth AAAI conference on artificial intelligence, 20,954.
45. Wang, Z. & Oates, T. (2015b). Imaging time-series to improve classification and imputation. arXiv preprint arXiv:1506.00327.
46. Zhi, W., Appling, A. P., Golden, H. E., Podgorski, J. & Li, L. (2024). Deep learning for water quality. Nature water, 2 (3), 228–241.
47. Zhu, Y., Brettin, T., Xia, F., Partin, A., Shukla, M., Yoo, H., Evrard, Y. A., Doroshow, J. H. & Stevens, R. L. (2021). Converting tabular data into images for deep learning with convolutional neural networks. Scientific reports, 11 (1), 11325.