Keywords:
transfer learning
computer vision
real-time processing
fine-tuning
automated grading

  • Abstract

    This study aimed to develop and evaluate the performance of deep learning models for classifying the quality of Golden Nam Dok Mai mangoes under real-world environmental conditions. A dataset of 1,200 mango images, categorized into three quality levels—Grade A, B, and C—was utilized. The experiments were conducted using 5-fold cross-validation across five distinct neural network architectures: custom CNN, ResNet50, VGG16, MobileNetV2, and EfficientNetB0, while comparing performance under conditions with and without image augmentation. The experimental results revealed that the MobileNetV2 model achieved the most distinguished performance, with both accuracy and F1-score reaching a perfect 100% after the fine-tuning process. This was followed by ResNet50, which achieved an accuracy of 99.44%, while VGG16 and EfficientNetB0 also demonstrated robust classification capabilities. Although statistical analysis using one-way ANOVA indicated no significant difference among the models (p > 0.05), the average performance of MobileNetV2 remained the highest. Beyond classification precision, this study emphasized computational costs and latency for industrial applicability. To ensure measurement stability, each model underwent 100 warm-up iterations. The custom CNN exhibited the lowest latency at 5.48 ms per image, followed by VGG16 and ResNet50 with latencies of 12.73 ms and 36.54 ms, respectively. The MobileNetV2 recorded a latency of 42.64 ms, confirming their technical readiness for real-time processing in high-speed conveyor systems. Furthermore, the perfect results obtained from MobileNetV2 were validated for stability, suggesting that the model effectively captured discriminative physical visual features under the defined experimental conditions, without evident overfitting. These findings highlight the significant potential of transfer learning combined with fine-tuning in enhancing fruit grading standards and increasing the competitiveness of the Thai agricultural sector. Finally, this research serves as a fundamental framework for future advancements, including the integration of multimodal learning and the implementation of Edge AI for commercial use and smart farming solutions.

  • References

    1. Barman, U., Choudhury, R. D., Sahu, D., & Barman, G. G. (2020). Comparison of convolution neural networks for smartphone image-based real-time classification of citrus leaf disease. Computers and Electronics in Agriculture, 177, 105661. https://doi.org/10.1016/j.compag.2020.105661
    2. Fu, L., Gao, F., Wu, J., Li, R., Karkee, M., & Zhang, Q. (2020). Application of consumer RGB-D cameras for fruit detection and localization in field: A critical review. Computers and Electronics in Agriculture, 177, 105687. https://doi.org/10.1016/j.compag.2020.105687
    3. Gulzar, Y. (2022). Fruit image classification model based on MobileNetV2 with deep transfer learning technique. Sustainability, 15(3), 1906. https://doi.org/10.3390/su15031906
    4. Hossain, M. A., Sakib, S., Abdullah, H. M., & Arman, S. E. (2024). Deep learning for mango leaf disease identification: A vision transformer perspective. Heliyon, 10(17), e36361. https://doi.org/10.1016/j.heliyon.2024.e36361
    5. Iqbal, H. M. R., Hakim, M. S., & Bakar, N. A. (2022). Classification and grading of harvested mangoes using convolutional neural network. International Journal of Fruit Science, 22(1), 95–112. https://doi.org/10.1080/15538362.2021.2023069
    6. Khandelwal, S., Raut, A., Vyawahare, H., Theng, D., & Dhande, S. (2024). Optimizing performance in mango plant leaf disease classification through advanced machine learning techniques. Engineering, Technology & Applied Science Research, 14(6), 18476–18480. https://doi.org/10.48084/etasr.8220
    7. Kiran, P. R., Yeasin, M., Aradwad, P., Arunkumar, T. V., & Parray, R. A. (2024). Advancing mango quality assurance: Non-destructive detection of spongy tissue using visible near-infrared spectroscopy and machine learning classification. Food Packaging and Shelf Life, 34, 12023. https://doi.org/10.1002/fpf2.12023
    8. Koirala, A., Walsh, K. B., Wang, Z., & Anderson, N. (2020). Deep learning for mango (Mangifera indica) panicle stage classification. Agronomy, 10(1), 143. https://doi.org/10.3390/agronomy10010143
    9. Mahmud, B. U., Al Mamun, A., Hossen, M. J., Hong, G. Y., & Jahan, B. (2024). Light-weight deep learning model for accelerating the classification of mango-leaf disease. Emerging Science Journal, 8(1), 28–42. https://doi.org/10.28991/ESJ-2024-08-01-03
    10. Mienye, I. D., & Swart, T. G. (2024). A comprehensive review of deep learning: Architectures, recent advances, and applications. Information, 15(12), 755. https://doi.org/10.3390/info15120755
    11. Min, X., Ye, Y., Xiong, S., & Chen, X. (2025). Computer vision meets generative models in agriculture: Technological advances, challenges and opportunities. Applied Sciences, 15(14), 7663. https://doi.org/10.3390/app15147663
    12. Muzamal, M., Hussain, M., & De Wibowo, A. (2025). Non-destructive mango quality prediction using machine learning algorithms. Engineering Proceedings, 107(1), 116. https://doi.org/10.3390/engproc2025107116
    13. Nithya, R., Santhi, B., Manikandan, R., Rahimi, M., & Gandomi, A. H. (2022). Computer vision system for mango fruit defect detection. Foods, 11(21), 3483. https://doi.org/10.3390/foods11213483
    14. O’Brien, C., Falagán, N., Kourmpetli, S., Landahl, S., Terry, L. A., & Alamar, M. C. (2024). Non-destructive methods for mango ripening prediction: Visible and near-infrared spectroscopy (visNIRS) and laser Doppler vibrometry (LDV). Postharvest Biology and Technology, 212, 112878. https://doi.org/10.1016/j.postharvbio.2024.112878
    15. Rahman, M. M., Basar, M. A., Shinti, T. S., Khan, M. S. I., Babu, H. M. H., & Uddin, K. M. M. (2023). A deep CNN approach to detect and classify local fruits through a web interface. Array, 18, 100321. https://doi.org/10.1016/j.atech.2023.100321
    16. Ratha, A. K., Barpanda, N. K., Sethy, P. K., & Behera, S. (2024). Automated classification of Indian mango varieties using machine learning and MobileNetV2 deep features. Traitement du Signal, 41(2), 425–433. https://doi.org/10.18280/ts.410210
    17. Sawant, K., Shenoy, A. K. B., Pai, S., & Shirwaikar, R. D. (2025). Comprehensive analysis of fruit variety classification: Techniques, challenges, and application. Procedia Computer Science, 238, 260–268. https://doi.org/10.1016/j.procs.2025.04.260
    18. Shahi, T. B., Sitaula, C., Neupane, A., & Guo, W. (2022). Fruit classification using attention-based MobileNetV2 for industrial applications. PLOS ONE, 17(2), e0264586. https://doi.org/10.1371/journal.pone.0264586
    19. Tai, N. D., Lin, W. C., Trieu, N. M., & Thinh, N. T. (2024). Development of a mango-grading and -sorting system based on external features, using machine learning algorithms. Agronomy, 14(4), 831. https://doi.org/10.3390/agronomy14040831
    20. Tapia-Mendez, E., Cruz-Albarran, I. A., Tovar-Arriaga, S., & Morales-Hernandez, L. A. (2023). Deep learning-based method for classification and ripeness assessment of fruits and vegetables. Applied Sciences, 13(22), 12504. https://doi.org/10.3390/app132212504
    21. Worasawate, D., Sakunasinha, P., & Chiangga, S. (2022). Automatic classification of the ripeness stage of mango fruit using a machine learning approach. AgriEngineering, 4(1), 32–47. https://doi.org/10.3390/agriengineering4010003
    22. Wu, Y.-C., Hsu, H.-W., & Chen, S.-W. (2025). Mobile-based deep learning approach for multi-object Aiwen mango grade classification. Electronics, 14(16), 3188. https://doi.org/10.3390/electronics14163188
    23. Xiao, F., Wang, H., Xu, Y., & Zhang, R. (2023). Fruit detection and recognition based on deep learning for automatic harvesting: An overview and review. Agronomy, 13(6), 1625. https://doi.org/10.3390/agronomy13061625
    24. Zheng, B., & Huang, T. (2021). Mango grading system based on optimized convolutional neural network. Journal of Food Quality, 2021, 2652487. https://doi.org/10.1155/2021/2652487
Creative Commons License

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

How to cite

Somkantha, K., Phiewdam, W., kultangwattana, W., & Kasabai, P. (2026). Performance evaluation of deep learning models for quality classification of Golden Nam Dok Mai mangoes. Multidisciplinary Science Journal, (| Accepted Articles). Retrieved from https://malque.pub/ojs/index.php/msj/article/view/14524
  • Article viewed - 178