MobileNetV3 for Durian Seedling Variety Classification Based on Leaf Images with Sobel Edge Detection Preprocessing
DOI:
https://doi.org/10.30871/jaic.v10i3.13085Keywords:
Convolutional Neural Network, Deep Learning, Durian Leaf Classification, MobileNetV3Small, Sobel Edge Detection, Transfer LearningAbstract
Durian is a high-value horticultural commodity in Indonesia, yet variety identification based on leaf morphology remains largely manual and prone to subjective error. This study develops a deep learning-based classification system for durian leaf varieties using MobileNetV3Small, designed for practical deployment on mobile and edge computing devices in agricultural field settings to support real-time variety identification by farmers and agricultural practitioners. A dataset of 1,680 images across four durian varieties Bawor, Musang King, Duri Hitam, and Super Tembaga was constructed through direct field collection and public dataset acquisition, followed by augmentation via rotation and flipping techniques. Preprocessing incorporated background removal, cropping, resizing to 224×224 pixels, and Sobel edge extraction to enhance morphological features such as leaf veins and contours. Transfer learning from ImageNet weights was applied, with training conducted in two phases: a 30-epoch warm-up with frozen base layers, followed by 90-epoch fine-tuning of the 20 deepest layers selected to adapt high-level semantic features while preserving general low-level representations and avoiding catastrophic forgetting using cosine annealing and early stopping. The model achieved a validation accuracy of 98.42% and a macro average F1-score of 0.9842, with only 4 misclassifications out of 253 test images. All misclassifications occurred exclusively between the morphologically similar Bawor and Duri Hitam classes. Comparative analysis against eight prior studies using similar durian leaf classification tasks suggests that the proposed approach achieves competitive performance relative to studies employing larger architectures such as ResNet and ConvNeXt, though direct comparison on an identical dataset was not conducted and remains a direction for future work.
Downloads
References
[1] Y. Miftahuddin and F. Zaelani, “Perbandingan Metode Efficientnet-B3 dan Mobilenet-V2 Untuk Identifikasi Jenis Buah-buahan Menggunakan Fitur Daun,” 2022.
[2] C. Algemayel, D. Abou Jaoude, S. Talhouk, I. Issa, and C. Ghassibe, “Advances in machine learning models for plant species identification: A scoping review,” Feb. 01, 2026, Elsevier B.V. doi: 10.1016/j.ecoinf.2025.103464.
[3] S. Nassor, M. Mushthofa, and K. Priandana, “Deep learning model for detection and classification of banana diseases based on leaf images,” in IOP Conference Series: Earth and Environmental Science, Institute of Physics, 2024. doi: 10.1088/1755-1315/1359/1/012010.
[4] U. Kulsum and A. Cherid, “Penerapan Convolutional Neural Network Pada Klasifikasi Tanaman Menggunakan ResNet50,” SIMKOM, vol. 8, no. 2, pp. 221–228, Jul. 2023, doi: 10.51717/simkom.v8i2.191.
[5] C. Y. Kim, K. S. Um, and S. W. Heo, “A novel MobileNet with selective depth multiplier to compromise complexity and accuracy,” ETRI Journal, vol. 45, no. 4, pp. 666–677, Aug. 2023, doi: 10.4218/etrij.2022-0103.
[6] C. P. Lee, K. M. Lim, Y. X. Song, and A. Alqahtani, “Plant-CNN-ViT: Plant Classification with Ensemble of Convolutional Neural Networks and Vision Transformer,” Plants, vol. 12, no. 14, Jul. 2023, doi: 10.3390/plants12142642.
[7] A. Naufal Abiyyu and M. Rahardi, “Comparison of Transfer learning Models MobileNetV3-Large and EfficientNet-B0 for Rice Leaf Disease Classification,” 2026. [Online]. Available: http://jurnal.polibatam.ac.id/index.php/JAIC
[8] S. J. Elroy, F. Nurdiyansyah, G. Priyandoko, and K. Kunci, “3 RD MDP Student Conference (MSC) 2024 56 | Universitas Multi Data Palembang Klasifikasi Daun Durian Pada Citra Dalam Menentukan Jenis Menggunakan Convolutional Neutral Network,” 2024.
[9] D. Kurniawan and D. Ariatmanto, “Identifikasi Varietas Bibit Durian Menggunakan Mobilenetv2 Berdasarkan Gambar Daun,” 2024. [Online]. Available: http://e-journal.stmiklombok.ac.id/index.php/jireISSN.2620-6900
[10] R. Klangbunrueang, W. Chansanam, N. Iam-On, and T. Boongoen, “AI-Driven Durian Leaf Disease Classification Using Benchmark CNN Architectures for Precision Agriculture,” Applied Sciences, vol. 16, no. 9, p. 4062, Apr. 2026, doi: 10.3390/app16094062.
[11] D. Diana et al., “Convolutional Neural Network Based Deep Learning Model for Accurate Classification of Durian Types,” Journal of Applied Data Sciences, vol. 6, no. 1, pp. 101–114, Jan. 2025, doi: 10.47738/jads.v6i1.480.
[12] N. M. Voo, T. M. Lim, and Y. M. Lim, “Fine-Tuning MobileNet for Durian Variety Classification,” in IEEE ICCBE 2025, Basel Switzerland: MDPI, Mar. 2026, p. 46. doi: 10.3390/engproc2026128046.
[13] P. Eiamin, P. Janoum, J. Puangmanee, and P. Chophuk, “Comparative Evaluation of YOLOv5, YOLOv8, and TensorFlow for Optimal Model Selection in Durian Leaf Disease Detection via Mobile Application,” in Frontiers in Artificial Intelligence and Applications, IOS Press BV, Mar. 2025, pp. 401–411. doi: 10.3233/FAIA241595.
[14] I. Wirabowo and I. Susilawati, “Implementasi Convolution Neural Network (CNN) untuk Deteksi Penyakit pada Daun Jagung Berbasis Citra Digital,” Jurnal Pustaka Data (Pusat Akses Kajian Database, Analisa Teknologi, dan Arsitektur Komputer), vol. 5, no. 1, pp. 233–241, Jun. 2025, doi: 10.55382/jurnalpustakadata.v5i1.1046.
[15] S. A. Wagle, R. Harikrishnan, S. H. M. Ali, and M. Faseehuddin, “Classification of plant leaves using new compact convolutional neural network models,” Plants, vol. 11, no. 1, Jan. 2022, doi: 10.3390/plants11010024.
[16] H. Pebrian and E. Hartati, “Potato Leaf Disease Classification Using MobileNetV3 Architecture with Adam and Stochastic Gradient Descent Optimizers,” Research Article, vol. 6, no. 1, pp. 81–95, 2026, doi: 10.53623/gisa.v6i1.1063.
[17] P. Sridhara Acharya, “MobileNetV3Small-CM FusionNet: A Lightweight Deep Learning Framework for Multi-Class Arecanut Disease Classification Using Feature Fusion,” 2024. [Online]. Available: https://www.jisem-journal.com/
[18] K. Navinkumar, R. Logesh, P. VishnuBabu, and A. V. Ananthalakshmi, “FPGA implementation of sobel edge detection algorithm,” EAI Endorsed Transactions on Internet of Things, vol. 10, 2024, doi: 10.4108/eetiot.5148.
[19] K. Alomar, H. I. Aysel, and X. Cai, “Data Augmentation in Classification and Segmentation: A Survey and New Strategies,” J. Imaging, vol. 9, no. 2, Feb. 2023, doi: 10.3390/jimaging9020046.
[20] Y. Apriansyah et al., “Implementation Of Edge Detection Using The Sobel Operator On Papaya Leaf Images,” Jurnal Ilmiah Informatika dan Komputer (Informatech), vol. 2, no. 2, pp. 164–168, 2025, doi: 10.69533.
[21] S. Farhadpour, T. A. Warner, and A. E. Maxwell, “Selecting and Interpreting Multiclass Loss and Accuracy Assessment Metrics for Classifications with Class Imbalance: Guidance and Best Practices,” Remote Sens. (Basel)., vol. 16, no. 3, Feb. 2024, doi: 10.3390/rs16030533.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Rusly Huda, Indah Susilawati
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License (Attribution-ShareAlike 4.0 International (CC BY-SA 4.0) ) that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
