PENERAPAN ARSITEKTUR EFFICIENTNET-B3 BERBASIS TRANSFER LEARNING UNTUK KLASIFIKASI PENYAKIT DAUN POHON KARET

ubang ubang; Asril Adi Sunarto; Fathia Frazna Az-Zahra

doi:10.23960/jitet.v14i3.10854

ubang ubang
Universitas muhammadiyah sukabumi prodi teknik informatika
Asril Adi Sunarto
Universitas muhammadiyah sukabumi
Fathia Frazna Az-Zahra
Universitas muhammadiyah sukabumi

DOI: https://doi.org/10.23960/jitet.v14i3.10854

Keywords EfficientNet-B3, Klasifikasi Citra, Penyakit Daun Karet, Transfer Learning, CRISP-DM

Abstract Views (Last 12 Months)

11 Abstract Views

10 Downloads

Abstract

Abstrak. Sektor perkebunan karet terancam oleh penyakit daun utama (Oidium, Corynespora, Colletotrichum) yang menurunkan produksi lateks. Identifikasi visual manual di lapangan memiliki keterbatasan kecepatan dan bersifat subjektif. Penelitian ini mengimplementasikan model Deep Learning arsitektur EfficientNet-B3 dengan Transfer Learning untuk klasifikasi penyakit daun karet berdasarkan metodologi CRISP-DM. Dataset terdiri dari 2.972 citra yang terbagi merata dalam empat kelas (Sehat, Colletotrichum, Corynespora, Oidium). Data diproses melalui penyesuaian dimensi 300x300 piksel, normalisasi, dan augmentasi. Hasil pengujian menggunakan Confusion Matrix pada 448 citra uji independen menghasilkan tingkat akurasi keseluruhan sebesar 99,11% dengan nilai presisi rata-rata 0,99. Evaluasi kompleksitas menunjukkan model ini sangat ringan dengan waktu inferensi 80,39 ms per citra. Model ini diintegrasikan ke dalam prototipe aplikasi web lokal menggunakan arsitektur hibrida FastAPI dan PHP yang dilengkapi Hybrid Validation System. Hasil integrasi ini menghasilkan sistem deteksi penyakit daun karet yang akurat dan efisien untuk lingkungan komputasi terbatas.

Abstract. The rubber plantation sector is threatened by major leaf diseases (Oidium, Corynespora, Colletotrichum) which decrease latex production. Manual visual identification in the field has speed limitations and is subjective. This study implements the EfficientNet-B3 Deep Learning architecture with Transfer Learning for rubber leaf disease classification based on the CRISP-DM methodology. The dataset consists of 2,972 images evenly divided into four classes (Healthy, Colletotrichum, Corynespora, Oidium). The data is processed through 300x300 pixel resizing, normalization, and augmentation. Testing using a Confusion Matrix on 448 independent test images yielded an overall accuracy of 99.11% with an average precision value of 0.99. Complexity evaluation shows this model is very lightweight with an inference time of 80.39 ms per image. This model is integrated into a local web application prototype using a hybrid architecture of FastAPI and PHP equipped with a Hybrid Validation System. This integration produces a highly accurate and efficient rubber leaf disease detection system for limited computing environments.

Downloads

Download data is not yet available.

References

S. Sagaff, S. Aliya, N. S. Ali, M. M. Mahyudin, W. Mui, and S. Buloh, “Emerging and Existing Major Leaf Diseases of Hevea brasiliensis in Malaysia 2 Integrated Pest and Disease Management Unit Experimental Station , Malaysian Rubber Board ,” vol. 3, no. 1, pp. 34–47, 2022.

T. R. Febbiyanti and R. Tistama, “Physiological Conditions and Production of Rubber Plant Affected by Pestalotiopsis with Fertilizer and Fungicide Application,” J. Penelit. Karet, vol. 39, no. 2, pp. 139–150, 2021.

A. I. C. Nyaka Ngobisa, M. A. Zainal Abidin, M. Y. Wong, and M. W. D. Wan Noordin, “Neofusicoccum ribis associated with leaf blight on rubber (Hevea brasiliensis) in Peninsular Malaysia,” Plant Pathol. J., vol. 29, no. 1, pp. 10–16, 2023, doi: 10.5423/PPJ.OA.07.2012.0110.

J. Menezes, A. Aasaithambi, D. Sam, G. Maheswari, and R. Santhosh, “Plant Leaf Disease Detection Using Deep Learning,” Proc. Int. Conf. Sustain. Comput. Integr. Commun. Chang. Landsc. AI, ICSCAI 2024, pp. 1–24, 2024, doi: 10.1109/ICSCAI61790.2024.10867022.

S. F. N. Nur Azizah and R. Fikri, “Analisis Perbandingan ALBERT, BERT, dan IndoBERT untuk Deteksi Ujaran Kebencian,” J. Pengolah. Bhs. Alami, vol. 4, no. 1, pp. 50–58, 2023.

J. Lu, L. Tan, and H. Jiang, “Review on convolutional neural network (CNN) applied to plant leaf disease classification,” Agric., vol. 11, no. 8, pp. 1–18, 2021, doi: 10.3390/agriculture11080707.

A. Asrafil, A. Paliwang, M. R. D. Septian, M. Cahyanti, and R. Swedia, “Klasifikasi Penyakit Tanaman Apel Dari Citra Daun Dengan,” Sebatik, no. 1410–3737, pp. 207–212, 2020.

S. M. Hassan, A. K. Maji, M. Jasiński, Z. Leonowicz, and E. Jasińska, “Identification of Plant-Leaf Diseases Using CNN and Transfer-Learning Approach,” Electron., vol. 10, no. 12, p. 1388, 2021.

Ü. Atila, M. Uçar, K. Akyol, and E. Uçar, “Plant leaf disease classification using EfficientNet deep learning model,” Ecol. Inform., vol. 61, p. 101182, 2021, doi: https://doi.org/10.1016/j.ecoinf.2020.101182.

S. S. Dash, S. K. Nayak, and D. Mishra, “A review on machine learning algorithms,” Smart Innov. Syst. Technol., vol. 153, pp. 495–507, 2021, doi: 10.1007/978-981-15-6202-0_51.

Y. Lecun, Y. Bengio, and G. Hinton, “Deep learning,” Nature, vol. 521, no. 7553, pp. 436–444, 2021, doi: 10.1038/nature14539.

E. Anggiratih, S. Siswanti, S. K. Octaviani, and A. Sari, “Klasifikasi Penyakit Tanaman Padi Menggunakan Model Deep Learning Efficientnet B3 dengan Transfer Learning,” J. Ilm. SINUS, vol. 19, no. 1, p. 75, 2021, doi: 10.30646/sinus.v19i1.526.

M. Liu, H. Liang, and M. Hou, “Research on cassava disease classification using the multi-scale fusion model based on EfficientNet and attention mechanism,” Front. Plant Sci., vol. 13, no. December, pp. 1–11, 2022, doi: 10.3389/fpls.2022.1088531.

S. O. Ngesthi and L. A. Wulandhari, “Cassava Diseases Classification using EfficientNet Model with Imbalance Data Handling,” J. Online Inform., vol. 9, no. 2, pp. 148–158, 2024, doi: 10.15575/join.v9i2.1300.

M. Iman, H. R. Arabnia, and K. Rasheed, “A Review of Deep Transfer Learning and Recent Advancements,” Technologies, vol. 11, no. 2, 2023, doi: 10.3390/technologies11020040.

T. Kattenborn, J. Leitloff, F. Schiefer, and S. Hinz, “Cnn5.Pdf,” vol. 173, no. November 2020, pp. 24–49, 2021.

R. Akyas Hifdzi Rahman, A. Adi Sunarto, and A. Asriyanik, “Penerapan You Only Look Once (Yolo) V8 Untuk Deteksi Tingkat Kematangan Buah Manggis,” JATI (Jurnal Mhs. Tek. Inform., vol. 8, no. 5, pp. 10566–10571, 2024, doi: 10.36040/jati.v8i5.10979.

G. Mahalaxmi, A. D. Donald, and T. A. S. Srinivas, “A Short Review of Python Libraries and Data Science Tools,” South Asian Res. J. Eng. Technol., vol. 5, no. 1, pp. 1–5, 2023, doi: 10.36346/sarjet.2023.v05i01.001.

O. E. Oghama, V. I. Omorusi, and E. S. Osazuwa, “Selected Leaf Diseases of Rubber: Symptoms and Control – A Review,” J. Appl. Sci. Environ. Manag., vol. 27, no. 11, pp. 2369–2373, 2023, doi: 10.4314/jasem.v27i11.1.

B. Tugrul, E. Elfatimi, and R. Eryigit, “Convolutional Neural Networks in Detection of Plant Leaf Diseases: A Review,” Agric., vol. 12, no. 8, 2022, doi: 10.3390/agriculture12081192.

A. Wongpanich et al., “Training EfficientNets at Supercomputer Scale: 83% ImageNet Top-1 Accuracy in One Hour,” 2021 IEEE Int. Parallel Distrib. Process. Symp. Work. IPDPSW 2021 - conjunction with IEEE IPDPS 2021, pp. 947–950, 2021, doi: 10.1109/IPDPSW52791.2021.00146.

A. C. Milano, “Klasifikasi Penyakit Daun Padi Menggunakan Model Deep Learning Efficientnet-B6,” J. Inform. dan Tek. Elektro Terap., vol. 12, no. 1, 2024, doi: 10.23960/jitet.v12i1.3855.