Implementasi Algoritma CNN dengan Arsitektur MobileNet untuk Klasifikasi Citra Daun Herbal
DOI:
https://doi.org/10.30865/jurikom.v13i1.9444Keywords:
Classification, Herbal Leaves, Deep Learning, CNN, MobileNetAbstract
Indonesia's diversity is very rich, one of which is that Indonesia has herbal plants which people believe are natural medicines for curing diseases. Herbal leaves show different variations in size and shape for each type, indicating that each leaf has special characteristics, shape, texture and size. Researchers used one of the Deep Learning methods, namely Convolutional Neural Network (CNN) for classifying herbal leaves. The field of image classification has found CNNs to be quite effective for image classification. CNN is a type of neural network with convolutional layers that has the ability to automatically extract important features from images. MobileNet is a CNN structure created by Google. MobileNet has advantages in efficient use of computing resources. Specifically, in the MobileNet network model, an attention module was added to improve the model's ability to extract more detailed image features, and dropout technology was added to prevent overfitting. This research method includes image preprocessing, training a convolutional neural network-based model, and evaluating its performance using accuracy, precision, recall, and F1 score metrics. Training was conducted for 20 epochs, and testing was conducted using data separated from the training data. The evaluation results show that the MobileNet model has the ability to extract visual features and produce herbal leaf image classification with an accuracy rate of 97.50% and precision, recall, and F1 scores of 98% each. The proposed model can be used in mobile-based herbal leaf identification applications due to its high performance and lightweight architecture. The stable accuracy curve at the final epoch indicates that the model does not experience significant overfitting and is able to generalize well to the test data
References
[1] F. Hidayanto and D. S. Ardi, “Tanaman herbal sebagai tanaman hias dan tanaman obat,” J. Inov. dan Kewirausahaan, vol. 4, no. 1, pp. 1–4, 2015.
[2] S. Siahaan and N. K. Aryastami, “Studi Kebijakan Pengembangan Tanaman Obat di Indonesia,” Media Penelit. dan Pengemb. Kesehat., vol. 28, no. 3, pp. 157–166, 2018, doi: 10.22435/mpk.v28i3.119.
[3] E. R. Pribadi, “Pasokan Dan Permintaan Tanaman Obat Indonesia Serta Arah Penelitian Dan Pengembangannya (Supply And Demand Of Indonesian Medicinal Plants And The Research And Development Direction),” Perspektif, vol. 8, no. 1, pp. 52–64, 2017, [Online]. Available: http://ejurnal.litbang.pertanian.go.id
[4] S. P. Backar, P. Purnawansyah, H. Darwis, and W. Astuti, “Hybrid Fourier Descriptor Naïve Bayes dan CNN pada Klasifikasi Daun Herbal,” J. Inform. J. Pengemb. IT, vol. 8, no. 2, pp. 126–133, 2023, doi: 10.30591/jpit.v8i2.5186.
[5] L. P. R. Noviana and I. N. B. S. Nugraha, “Perbandingan Klasifikasi Citra Daun Herbal Menggunakan Metode Logistic Regression dan Decision Tree Classifier Berdasarkan Fitur (Warna, GLCM, Bentuk),” JITU J. Inform. Technol. Commun., vol. 7, no. 2, pp. 126–133, 2023, doi: 10.36596/jitu.v7i2.1241.
[6] Adela Regita Azzahra, “Klasifikasi Daun Herbal Menggunakan Metode CNN dan Naïve Bayes dengan Fitur GLCM,” Indones. J. Comput. Sci., vol. 12, no. 4, 2023, doi: 10.33022/ijcs.v12i4.3362.
[7] S. Baba, P. Kumari, and P. Saxena, “Retinal Disease Classification Using Custom CNN Model from OCT Images,” Procedia Comput. Sci., vol. 235, no. 2023, pp. 3142–3152, 2024, doi: 10.1016/j.procs.2024.04.297.
[8] T. Bariyah, M. A. Rasyidi, and N. Ngatini, “Convolutional Neural Network untuk Metode Klasifikasi Multi-Label pada Motif Batik,” Techno.Com, vol. 20, no. 1, pp. 155–165, 2021, doi: 10.33633/tc.v20i1.4224.
[9] Y. Hao, “Convolutional Neural Networks for Image Classification,” Proc. - 2021 2nd Int. Conf. Artif. Intell. Comput. Eng. ICAICE 2021, vol. 8, no. 5, pp. 342–345, 2021, doi: 10.1109/ICAICE54393.2021.00073.
