Prediksi Perbandingan Kinerja MobileNetV2 dan EfficientNetB0 dalam Prediksi Kematangan Buah Jaboticaba

Mahardika Wildan Fitrian; Wiwi Widayani; Heri Sismoro; Nur Aini; Dwi Nurani

doi:10.47134/jacis.v6i1.177

Authors

Mahardika Wildan Fitrian Universitas Amikom Yogyakarta
Wiwi Widayani Universitas AMIKOM Yogyakarta https://orcid.org/0000-0003-4782-6523
Heri Sismoro Universitas AMIKOM Yogyakarta
Nur Aini Universitas Amikom Yogyakarta
Dwi Nurani Universitas Amikom Yogyakarta

DOI:

https://doi.org/10.47134/jacis.v6i1.177

Keywords:

Jaboticaba, deep learning, MobileNetV2, EfficientNetB0, prediksi kematangan

Abstract

Penentuan kematangan buah yang akurat penting untuk meningkatkan kualitas panen dan efisiensi distribusi, namun praktik manual masih rentan terhadap subjektivitas. Penelitian ini membandingkan kinerja Convolutional Neural Network (CNN), MobileNetV2, dan EfficientNetB0 dalam memprediksi jumlah hari menuju kematangan buah Jaboticaba (Plinia cauliflora) berbasis citra RGB. Dataset diperoleh dari pengamatan lapangan dan diuji dalam dua skenario: 656 citra dan dataset hasil augmentasi sebanyak ±2.500 citra. Seluruh model dilatih dengan optimizer Adam dan dievaluasi menggunakan Mean Absolute Error (MAE). Pada dataset kecil, MobileNetV2 menunjukkan performa terbaik dengan MAE uji 1,54 hari, lebih baik dibanding CNN (3,41 hari) dan EfficientNetB0 (6,05 hari). Setelah jumlah data diperbesar, kinerja seluruh model meningkat, dengan MobileNetV2 tetap unggul (MAE 0,95 hari), diikuti EfficientNetB0 (1,30 hari) dan CNN (1,86 hari). Peningkatan terbesar terjadi pada EfficientNetB0, yang mengindikasikan bahwa model berkapasitas tinggi membutuhkan data lebih besar untuk belajar secara optimal. Hasil ini menegaskan bahwa ukuran dataset berpengaruh signifikan terhadap akurasi, serta bahwa MobileNetV2 memberikan keseimbangan terbaik antara ketepatan dan stabilitas

References

[1] Y. LeCun, Y. Bengio, and G. Hinton, “Deep learning,” Nature, vol. 521, no. 7553, pp. 436–444, 2015, doi: 10.1038/nature14539. DOI: https://doi.org/10.1038/nature14539

[2] U. Nawaz et al., “AI in Agriculture: A Survey of Deep Learning Techniques for Crops, Fisheries and Livestock,” arXiv:2507.22101, 2025.

[3] M. Naghipour, L. Sook Ling, and T. Connie, “A Review of AI Techniques in Fruit Detection and Classification: Analyzing Data, Features and AI Models Used in Agricultural Industry,” IJTech, vol. 15, no. 3, 2024. DOI: https://doi.org/10.14716/ijtech.v15i3.6404

[4] F. Xiao, X. Zhang, and Y. Li, “Fruit Detection and Recognition Based on Deep Learning for Automatic Harvesting: An Overview and Review,” Agronomy, vol. 13, no. 6, 2023, doi: 10.3390/agronomy13061579. DOI: https://doi.org/10.3390/agronomy13061625

[5] U. Amin et al., “Automatic Fruits Freshness Classification Using CNN and Transfer Learning,” Applied Sciences, vol. 13, no. 14, 2023, doi: 10.3390/app13148254. DOI: https://doi.org/10.3390/app13148087

[6] H. Mundhada et al., “Fruit Detection and Three-Stage Maturity Grading Using CNN,” International Journal of Next-Generation Computing, vol. 14, no. 1, 2023. DOI: https://doi.org/10.47164/ijngc.v14i1.1099

[7] V. Lusen, I. Purnamasari, and B. N. Sari, “Penerapan CNN untuk Klasifikasi Tingkat Kematangan Berry dengan Augmentasi Transformasi dan ColorJitter Menggunakan TensorFlow,” Jurnal Lentera, vol. 23, no. 3, pp. 418–427, 2024.

[8] B. Benmouna et al., “Convolutional Neural Networks for Estimating the Ripening State of Fuji Apples Using Visible and Near-Infrared Spectroscopy,” Food and Bioprocess Technology, vol. 15, pp. 2226–2236, 2022, doi: 10.1007/s11947-022-02891-5. DOI: https://doi.org/10.1007/s11947-022-02880-7

[9] A. R. Rachmad Dianto et al., “Analisis Efektivitas Algoritma MobileNetV3-Large dan EfficientNet-B0 untuk Klasifikasi Citra Penyakit Daun Jeruk,” Jurnal Informatika dan Teknik Elektro Terapan, vol. 13, no. 3, 2025. DOI: https://doi.org/10.23960/jitet.v13i3.6956

[10] A. R. R. Ivani et al., “Perbandingan VGG16 dan MobileNetV2 untuk Klasifikasi Tingkat Kematangan Buah Apel,” Jurnal Sistem Komputer Asia, vol. 3, no. 1, 2025.

[11] M. Irvai and D. Mahdalena, “Penerapan Teknik Masking dan Augmentasi pada Arsitektur CNN dan MobileNetV2 untuk Klasifikasi Fase Pertumbuhan Tomat,” JTIUST, vol. 10, no. 1, 2025.

[12] M. Han and C. Yi, “Deep Convolutional Neural Networks for Palm Fruit Maturity Classification,” arXiv:2502.20223, 2025.

[13] J. Schmidhuber, “Deep learning in neural networks: An overview,” Neural Networks, vol. 61, pp. 85–117, 2015, doi: 10.1016/j.neunet.2014.09.003. DOI: https://doi.org/10.1016/j.neunet.2014.09.003

[14] M. Tan and Q. Le, “EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks,” in Proc. ICML, 2019.

[15] Sandler, M., Howard, A., Zhu, M., Zhmoginov, A., & Chen, L.-C, “MobileNetV2: Inverted Residuals and Linear Bottlenecks,” Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 4510–4520, 2018. DOI: https://doi.org/10.1109/CVPR.2018.00474

[16] Sahil, “Impact of Dataset Size on Deep Learning Model,“impact-of-dataset-size-on-deep-learning-model, [Online]. Tersedia : https://www.geeksforgeeks.org/deep-learning/impact-of-dataset-size-on-deep-learning-model/, [Diakses : 28 April 2026]

[17] Ali Hasan Md. Linkon, Md. Mahir Labib, Faisal Haque Bappy, Soumik Sarker, Marium-E-Jannat, Md Saiful Islam, “Deep Learning Approach Combining Lightweight CNN Architecture with Transfer Learning: An Automatic Approach for the Detection and Recognition of Bangladeshi Banknotes, “ 11th International Conference on Electrical and Computer Engineering (ICECE), 2020.