Yenilebilen Yabani Bitki Görüntülerinin Derin Öğrenme Tabanlı Sınıflandırılması: Mobil Uygulama Örneği

Year 2024, Volume: 2 Issue: 2, 17 - 32, 31.08.2024

Mohamad Alhaj Rabia , İrem Nur Ecemiş , Mustafa Karhan

Abstract

Bu çalışmada, yenilebilir yabani bitkilerin (YYB) tanımlanması amacıyla geliştirilen bir mobil uygulama tanıtılmaktadır. MobileNetV2 mimarisi kullanılarak oluşturulan derin öğrenme modeli, 35 farklı bitki türünü tanımlayabilmektedir. Model, toplamda 16.500 görüntüden oluşan geniş bir veri setiyle eğitilmiştir. Eğitim sürecinde veri arttırma teknikleri kullanılarak modelin genelleme yeteneği geliştirilmiştir. Bu teknikler, görüntülerin döndürülmesi, kaydırılması, yakınlaştırılması ve yatay olarak çevrilmesi gibi çeşitli işlemleri içermektedir. Flask API aracılığıyla entegre edilen model, React Native ile geliştirilen mobil uygulama üzerinden kullanılabilir hale getirilmiştir. Uygulama, kullanıcıların bitkiler hakkında bilgi edinmelerini ve bu bilgileri favorilerine eklemelerine imkan sağlar. Ek olarak, kullanıcıların önceki taramalarını ve favori bitkilerini listelemelerine olanak tanır. Geliştirilen sistem, kullanıcı dostu arayüzü ve yüksek doğruluk oranıyla YYB’nin tanımlanmasında etkili bir çözüm sunmaktadır. Elde edilen sonuçlar, modelin eğitim doğruluğunun %85 ve doğrulama doğruluğunun %82 olduğunu göstermektedir. Bu çalışma, mobil cihazlar üzerinden YYB tanımlama alanında önemli bir adım atmakta ve doğa meraklıları, botanikçiler ve araştırmacılar için değerli bir araç sunmaktadır. Geliştirilen uygulama, benzer çalışmalara kıyasla daha geniş bir veri seti ve yüksek doğruluk oranı ile dikkat çekmektedir.

Keywords

Yenilebilen Yabani Bitkiler, Derin Öğrenme, MobileNetV2, React Native

Deep Learning Based Classification of Edible Wild Plant Images: A Mobile Application Model

Abstract

In this study, a mobile application developed for the identification of edible wild plants (YYB) is introduced. The deep learning model created using the MobileNetV2 architecture can identify 35 plant species. The model was trained with a large dataset consisting of 16,500 images in total. The model's generalization ability was increased by using data augmentation techniques during the training process. These techniques include various operations such as rotating, panning, zooming, and horizontally flipping the images. The model, integrated via the Flask API, was made available via the mobile application developed with React Native. The application allows users to obtain plant information and add it to their favorites. In addition, it allows users to list their previous scans and favorite plants. The developed system offers an effective solution for identifying YYB with its user-friendly interface and high accuracy rate. The results show that the model's training accuracy is 85%, and the validation accuracy is 82%. This study takes an essential step in identifying YYB via mobile devices and provides a valuable tool for nature enthusiasts, botanists, and researchers. The developed application attracts attention with its larger data set and higher accuracy rate than similar studies.

Keywords

Edible Wild Plants, Deep Learning, MobileNetV2, React Native

References

• Shaheen, S., Ahmad, M., Haroon, N., "Edible wild plants: a solution to overcome food insecurity," Edible Wild Plants: An Alternative Approach to Food Security 41-57 (2017). Doi: 10.1007/978-3-319-63037-3.
• Teklehaymanot, T., Giday, M., "Ethnobotanical study of wild edible plants of Kara and Kwego semi-pastoralist people in Lower Omo River Valley, Debub Omo Zone, SNNPR, Ethiopia," Journal of Ethnobiology and Ethnomedicine 6:1 1-8 (2010). Doi: 10.1186/1746-4269-6-23.
• Jamdhade, V. M., "Wild edible plants used by the tribes of Panvel and Uran Tahsils in Alibaugh District, India: Ethnobotanical application and tribal recipes," Journal of Botanical Research 4:1 13-19 (2022). Doi: 10.30564/jbr.v4i1.4280.
• Ojelel, S., Kakudidi, E. K., "Wild edible plant species utilized by a subsistence farming community in Obalanga sub-county, Amuria district, Uganda," Journal of Ethnobiology and Ethnomedicine 11:1 1-8 (2015). Doi: 10.1186/s13002-015-0037-7.
• Kohila, A., Mary Kensa, V., "Survey of wild edible plants of Dhanakarkulam Panchayath, Tirunelveli District, Tamil Nadu, India," Kongunadu Research Journal 6:2 20-27 (2019). Doi: 10.26524/krj297.
• Kebede, A., Tesfaye, W., Fentie, M., Zewide, H., "An ethnobotanical survey of wild edible plants commercialized in Kefira Market, Dire Dawa City, eastern Ethiopia," Plant 5:2 42-46 (2017). Doi: 10.11648/j.plant.20170502.13.
• Panda, S. P., Mazhar, Z., Chakraborty, K., Dasgupta, S., Kamila, P. K., Hameed, S. S., Sharief, M. U., "Diversity of wild edible fruit plants of Acharya Jagadish Chandra Bose Indian Botanic Garden, Howrah, West Bengal, India," Plant Archives 24:1 1463-1472 (2024).
• Gautam, R. S., Shrestha, S. J., Shrestha, I., "Wild edible fruits of Nepal," International Journal of Applied Sciences and Biotechnology 8:3 289-304 (2020). Doi: 10.3126/ijasbt.v8i3.31561.
• Addis, G., Urga, K., Dikasso, D., "Ethnobotanical study of edible wild plants in some selected districts of Ethiopia," Human Ecology 33:1 83-118 (2005). Doi: 10.1007/s10745-005-1656-0.
• Partridge, R. "Wild Edible Plants," Kaggle, 2024. Erişim: 27 Ağustos 2024. [Çevrimiçi]. Erişim adresi: https://www.kaggle.com/datasets/ryanpartridge01/wild-edible-plants.
• Partridge, R. "Wild Edible Plant Classifier," Github, 2024. Erişim: 27 Ağustos 2024. [Çevrimiçi]. Erişim adresi: https://github.com/Achronus/wep-classifier.
• Yang, S. Xiao, W. Zhang, M. Guo, S. Zhao, J. and Shen, F. "Image Data Augmentation for Deep Learning: A Survey," arXiv, 2023. Erişim: 27 Ağustos 2024. [Çevrimiçi]. Erişim adresi: https://arxiv.org/abs/2204.08610.
• Tatar, A., Haghighi, M., Zeinijahromi, A., "Experiments on image data augmentation techniques for geological rock type classification with convolutional neural networks," Journal of Rock Mechanics and Geotechnical Engineering (2024). Doi: 10.1016/j.jrmge.2024.02.015.
• Rahmatullah, P. Abidin, T. F. Misbullah, A. Nazaruddin, A. “Effectiveness of Data Augmentation in Multi-class Face Recognition,” 2021 5th International Conference on Informatics and Computational Sciences (ICICoS), 64-68 (2021). Doi: 10.1109/ICICoS52554.2021.9651780.
• Yang L., Hanneke S., Carbonell J., “A theory of transfer learning with applications to active learning,” Machine Learning 90 161-189 (2013). Doi: 10.1007/s10994-012-5310-7..
• Pan S. J., “Transfer Learning,” Data Classification: Algorithms and Applications 21 537-570 (2014).
• Yong L., Ma L., Sun D., Du L., “Application of MobileNetV2 to waste classification,” PLOS ONE 18 1-16 (2023). Doi: 10.1371/journal.pone.0282336..
• Abadi M., Agarwal A., Barham P., Brevdo E., Chen Z., Citro C., Corrado G. S., Davis A., Dean J., Devin M., "TensorFlow: Large-scale machine learning on heterogeneous distributed systems," arXiv (2016). arXiv:1603.04467..
• Rashidi, M., "Application of TensorFlow Lite on embedded devices: A hands-on practice of TensorFlow model conversion to TensorFlow Lite model and its deployment on Smartphone to compare model’s performance," Dissertation 1-43 (2022).
• Verma G., Gupta Y., Malik A. M., Chapman B., “Performance evaluation of deep learning compilers for edge inference,” 2021 IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW) 858-865 (2021). Doi: 10.1109/IPDPSW52791.2021.00137..
• Vyshnavi V. R., Malik A., “Efficient way of web development using Python and Flask,” International Journal of Recent Research Aspects 6:2 16-19 (2019).
• Alemu M. B., “REST API: Implementation with Flask-Python,” Lisans, Lapland UAS, (2014).
• React Native, "React Native Documentation," [Çevrimiçi]. Available: https://reactnative.dev/. [Erişim: 27 Ağustos 2024].
• Fentaw A. E., “Cross platform mobile application development: a comparison study of React Native Vs Flutter,” Yüksek Lisans, University of Jyväskylä Faculty of Information Technology, (2020).
• Danielsson W., “React Native application development,” Linköpings universitet, Swedia 10:4 1-10 (2016).
• Hutri H., “Comparison of React Native and Expo,” Yüksek Lisans, Lappeenranta–Lahti University Software Engineering and Digital Transformation, (2023).
• Järveläinen, H., "Creating a React Native UI component Library," Lisans, Turku UAS Information and Communications Technology (2024).
• Krstinić D., Braović M., Šerić L., Božić-Štulić D., “Multi-label classifier performance evaluation with confusion matrix,” Computer Science & Information Technology 10 1-14 (2020). Doi: 10.5121/csit.2020.100801.
• Flach P., Kull M., “Precision-recall-gain curves: PR analysis done right,” Advances in Neural Information Processing Systems 28 1-9 (2015).
• Xie Y., Zhu C., Zhou W., Li Z., Liu X., Tu M., “Evaluation of machine learning methods for formation lithology identification: A comparison of tuning processes and model performances,” Journal of Petroleum Science and Engineering 160 182-193 (2018). Doi:10.1016/j.petrol.2017.10.028
• Lipton, Z. C., Elkan, C., Naryanaswamy, B., "Thresholding classifiers to maximize F1 score," arXiv (2014). arXiv:1402.1892.