Geometric Data Augmentation with a Two-Stage Fine-Tuning Strategy for EfficientNetB3-Based Fruit Condition Classification

Hana Sajida Azhurra; Sugiyarto Surono; Aris Thobirin

doi:10.30829/zero.v10i1.28735

Geometric Data Augmentation with a Two-Stage Fine-Tuning Strategy for EfficientNetB3-Based Fruit Condition Classification

Hana Sajida Azhurra, Sugiyarto Surono, Aris Thobirin

Abstract

Accurate fruit condition classification is essential for automated food safety assessment, particularly due to health risks associated with chemical contaminants such as formalin. However, reliable generalization in automated inspection systems remains challenging because limited visual variation in image datasets often leads to overfitting in deep learning models. To address this challenge, this study proposes an EfficientNetB3-based framework that integrates geometric data augmentation with a structured two-stage fine-tuning strategy to improve robustness and training stability. The proposed model achieved 99% test accuracy with consistent cross-dataset performance. The framework also demonstrated stable optimization behavior across training stages, indicating improved generalization capability. From a practical perspective, the proposed approach may support scalable food quality monitoring and automated sorting in agricultural supply chains, as well as preliminary food safety screening in large-scale inspection processes.

Keywords

Deep Learning; EfficientNetB3; Fruit Condition Classification; Geometric Data Augmentation; Two-Stage Fine-Tuning.

Full Text:

PDF

References

N. L. Rane, M. Paramesha, S. P. Choudhary, and J. Rane, “Machine Learning and Deep Learning for Big Data Analytics: A Review of Methods and Applications,” Partners Universal International Innovation Journal, vol. 2, no. 3, pp. 172–197, 2024, doi: 10.5281/zenodo.12271006.

Y. E. Nagaty, “Digital Image Analysis (DIA) in Food Technology: An Overview,” Alexandria Journal of Food Science and Technology, vol. 21, no. 2, pp. 29–34, 2024, doi: 10.21608/ajfs.2025.340104.1062.

A. Bhargava and A. Bansal, “Fruits and vegetables quality evaluation using computer vision: A review,” Journal of King Saud University — Computer and Information Sciences, vol. 33, no. 3, pp. 243–257, 2021, doi: 10.1016/j.jksuci.2018.06.002.

I. H. Sarker, “Deep learning: A comprehensive overview on techniques, taxonomy, applications and research directions,” SN Computer Science, vol. 2, no. 6, Art. no. 420, 2021, doi: 10.1007/s42979-021-00815-1.

F. Badri, M. T. Alawiy, and E. M. Yuniarno, “Deep learning architecture based on convolutional neural network (CNN) in image classification,” Jurnal Ilmiah KURSOR, vol. 12, no. 2, pp. 83–92, 2023, doi: 10.21107/kursor.v12i2.349.

M. W. Sari, S. P. Sitorus, Rohani, and R. Pane, “Implementation of Convolutional Neural Network (CNN) Method in Determining the Level of Ripeness of Mango Fruit Based on Image,” Jurnal Penelitian Pendidikan IPA, vol. 11, no. 5, pp. 419–428, 2025, doi: 10.29303/jppipa.v11i5.11436.

M. Leslie, J. Vermaat, L. Haak, et al., “A FoodSafeR Perspective on Emerging Food Safety Hazards and Associated Risks,” Frontiers in Sustainable Food Systems, vol. 9, 2025, Art. no. 1646792, doi: 10.3389/fsufs.2025.1646792.

U. Agustina, R. M. Sari, M. V. Humairo, and E. O. P. Dewi, “Food Safety Monitoring: Formaldehyde Health Risk Assessment on Imported Fruits in Indonesia 2014-2022,” Journal of Global Research in Public Health, vol. 8, no. 2, pp. 206–215, 2023, doi: 10.30994/jgrph.v8i2.477.

H. Xu, H. Chen, Y. Li, T. Luo, D. Zhao, X. Chen, H. Zhang, X. Hu, H. Xu, Y. Wang, Y. Shentu, and Z. Tong, “Dietary formaldehyde: a silent aggravator of diabetes and cognitive impairments,” Nutrition & Diabetes, vol. 15, Art. no. 35, 2025, doi: 10.1038/s41387-025-00390-x.

A. Sattar, M. A. M. Ridoy, A. K. Saha, H. M. Hasan Babu, and M. N. Huda, “Computer vision based deep learning approach for toxic and harmful substances detection in fruits,” Heliyon, vol. 10, no. 3, Art. no. e25371, 2024, doi: 10.1016/j.heliyon.2024.e25371.

Y. K. Anagaw, W. Ayenew, L. W. Limenh, D. T. Geremew, M. C. Worku, T. A. Tessema, W. Simegn, and M. L. Mitku, “Food adulteration: Causes, risks, and detection techniques—review,” SAGE Open Medicine, vol. 12, pp. 1–10, 2024, doi: 10.1177/20503121241250184.

J. Rodrigues, C. Saraiva, J. García-Díez, J. Castro, and A. Esteves, “Evaluating the Effectiveness of Food Safety Policies in Portugal: A Stakeholder-Based Analysis of Challenges and Opportunities for Food Safety Governance,” Foods, vol. 14, no. 9, Art. no. 1534, 2025, doi: 10.3390/foods14091534.

I. Rojas Santelices, P. Gupta, C. Mallor, and M. Chandra, “Artificial vision systems for fruit inspection and classification,” Sensors, vol. 25, no. 5, Art. no. 1524, 2025, doi: 10.3390/s25051524.

K. Shehzad, U. Ali, and A. Munir, “Computer Vision for Food Quality Assessment: Advances and Challenges,” Global Journal of Machine Learning and Computing, vol. 1, no. 1, pp. 76–92, 2025, doi: 10.70445/gjmlc.1.1.2025.76-92.

A. Salari, A. Djavadifar, X. Liu, and H. Najjaran, “Object recognition datasets and challenges: A review,” Neurocomputing, vol. 495, pp. 129–152, 2022, doi: 10.1016/j.neucom.2022.01.022.

H. K. Dishar and L. A. Muhammed, “A Review of the Overfitting Problem in Convolution Neural Network and Remedy Approaches,” Journal of Al-Qadisiyah for Computer Science and Mathematics, vol. 15, no. 2, pp. 155–165, 2023, doi: 10.29304/jqcm.2023.15.2.1240.

A. Larasati, S. Surono, A. Thobirin, and D. A. Dewi, “Performance analysis of resampling techniques for overcoming data imbalance in multiclass classification,” JUITA: Jurnal Informatika, vol. 13, no. 1, pp. 57–66, Mar. 2025, doi: 10.30595/juita.v13i1.25270.

T. Islam, M. S. Hafiz, J. R. Jim, M. M. Kabir, and M. F. Mridha, “A systematic review of deep learning data augmentation in medical imaging: Recent advances and future research directions,” Healthcare Analytics, vol. 5, Art. no. 100340, 2024, doi: 10.1016/j.health.2024.100340.

H. Yuan, M. Zhu, R. Yang, H. Liu, I. Li, and C. Hong, “Rethinking Domain-Specific Pretraining by Supervised or Self-Supervised Learning for Chest Radiograph Classification: A Comparative Study Against ImageNet Counterparts in Cold-Start Active Learning,” Health Care Science, vol. 4, no. 2, pp. 110–143, 2025, doi: 10.1002/hcs2.70009.

J. Howard and S. Ruder, “Universal language model fine-tuning for text classification,” in Proc. 56th Annu. Meet. Assoc. Comput. Linguistics (ACL), Melbourne, Australia, vol. 1, pp. 328–339, 2018, doi: 10.18653/v1/P18-1031.

H. Xu, S. Ebner, M. Yarmohammadi, A. S. White, B. Van Durme, and K. Murray, “Gradual fine-tuning for low-resource domain adaptation,” in Proc. 2nd Workshop Domain Adaptation for NLP (Adapt-NLP) at EACL, Kyiv, Ukraine, 2021, pp. 214–221.

M. Tan and Q. V. Le, “EfficientNet: Rethinking model scaling for convolutional neural networks,” in Proc. 36th Int. Conf. Mach. Learn. (ICML), Proc. Mach. Learn. Res., vol. 97, 2019, pp. 6105–6114.

Khadijah, R. Kusumaningrum, Rismiyati, and N. Sabilly, “EfficientNet Model for Multiclass Classification of The Correctness of Wearing Face Mask,” Indonesian Journal of Electrical Engineering and Informatics, vol. 13, no. 1, pp. 18–29, 2025, doi: 10.52549/ijeei.v13i1.5197.

A. Alshoraihy, “EfficientNetB3 in Leukemia Detection: Advancements in Medical Imaging Analysis,” Medinformatics, vol. 2, no. 3, pp. 219–225, 2025, doi: 10.47852/bonviewMEDIN52023293.

J. Sharma, “Advanced Fruit Classification Using EfficientNetB3 Focusing on Apples and Tomatoes,” in Proc. 3rd Int. Conf. Advancement Technol. (ICONAT), Goa, India, pp. 1–4, 2024, doi: 10.1109/ICONAT61936.2024.10774951.

S. S. Reka, A. Bagelikar, P. Venugopal, V. Ravi, and H. Devarajan, “Deep Learning-Based Classification of Rotten Fruits and Identification of Shelf Life,” Computers, Materials and Continua, vol. 78, no. 1, pp. 781–794, 2024, doi: 10.32604/cmc.2023.043369.

S. S. S. Palakodati, V. R. R. Chirra, Y. Dasari, and S. Bulla, “Fresh and rotten fruits classification using CNN and transfer learning,” Revue d’Intelligence Artificielle, vol. 34, no. 5, pp. 617–622, Oct. 2020, doi: 10.18280/ria.340512.

M. H. I. Bijoy, S. Z. Tasnim, S. A. Awsaf, and M. Z. Hasan, “FruitVision: A benchmark dataset for fresh, rotten, and formalin-mixed fruit detection,” Data in Brief, vol. 61, p. 111752, 2025, doi: 10.1016/j.dib.2025.111752.

S. Mohammadi, S. Sattarpanah Karganroudi, M. Adda, and H. Ibrahim, “Toward smart railway maintenance: AI-enhanced Non-Destructive Evaluation using Vision Transformers and CNNs for fastener defect detection,” Green Energy and Intelligent Transportation, vol. 5, p. 100332, 2026, doi: 10.1016/j.geits.2025.100332.

A. A. A. El-Aziz, M. A. Mahmood, and S. A. El-Ghany, “Enhancing Early Detection of Oral Squamous Cell Carcinoma: A Deep Learning Approach with LRT-Enhanced EfficientNet-B3 for Accurate and Efficient Histopathological Diagnosis,” Diagnostics, vol. 15, no. 13, Art. no. 1678, 2025, doi: 10.3390/diagnostics15131678.

C. Shorten and T. M. Khoshgoftaar, “A survey on Image Data Augmentation for Deep Learning,” Journal of Big Data, vol. 6, no. 60, 2019, doi: 10.1186/s40537-019-0197-0.

S. Park, J. Kim, S. Wang, and J. Kim, “Effectiveness of image augmentation techniques on non-protective personal equipment detection using YOLOv8,” Applied Sciences, vol. 15, no. 5, Art. no. 2631, 2025, doi: 10.3390/app15052631.

A. Buslaev, V. I. Iglovikov, E. Khvedchenya, A. Parinov, M. Druzhinin, and A. A. Kalinin, “Albumentations: Fast and flexible image augmentations,” Information, vol. 11, no. 2, Art. no. 125, 2020, doi: 10.3390/info11020125.

Henderi, T. Wahyuningsih, and E. Rahwanto, “Comparison of min-max normalization and z-score normalization in the k-nearest neighbor (kNN) algorithm to test the accuracy of types of breast cancer,” International Journal of Informatics and Information System, vol. 4, no. 1, pp. 13–20, 2021, doi: 10.47738/ijiis.v4i1.73.

A. Mamadmurodov, S. Umirzakova, M. Rakhimov, A. Kutlimuratov, Z. Temirov, R. Nasimov, A. Meliboev, A. Abdusalomov, and Y. I. Cho, “A Hybrid Deep Learning Model for Early Forest Fire Detection,” Forests, vol. 16, no. 5, Art. no. 863, 2025, doi: 10.3390/f16050863.

M. K. Anam, S. Defit, Haviluddin, L. Efrizoni, and M. B. Firdaus, “Early Stopping on CNN-LSTM Development to Improve Classification Performance,” Journal of Applied Data Sciences, vol. 5, no. 3, pp. 1175–1188, 2024, doi: 10.47738/jads.v5i3.312.

I. Goodfellow, Y. Bengio, and A. Courville, Deep Learning. Cambridge, MA, USA: MIT Press, 2016.

M. Grandini, E. Bagli, and G. Visani, “Metrics for multi-class classification: An overview,” arXiv:2008.05756, 2020, doi: 10.48550/arXiv.2008.05756.

J. Opitz, “A closer look at classification evaluation metrics and a critical reflection of common evaluation practice,” Transactions of the Association for Computational Linguistics, vol. 12, pp. 820–836, 2024, doi: 10.1162/tacl_a_00675.

J. Sharma and B. V. Kumar, “Automated classification of fresh and rotten fruits using ResNet50 for enhanced food quality control and waste reduction,” in Proc. Int. Conf. Pervasive Computational Technologies (ICPCT), Greater Noida, India, pp. 159–163, 2025, doi: 10.1109/ICPCT64145.2025.10939134.

DOI: http://dx.doi.org/10.30829/zero.v10i1.28735