Çiçek Görüntü Sınıflandırılmasında Ön Eğitimli Evrişimsel Sinir Ağlarının Performans Karşılaştırması

Öz

Çiçek sınıflandırması botonikten, ekolojik çalışmalara kadar birçok alan için önemlidir. Çiçek görüntülerinin net şekilde belirgin olmaması, yaprakların, dalların görüntüyü kapatması ve benzer özellikte çiçeklerin çok olması sınıflandırma çalışmalarında rastlanan zorluklardandır. Çalışmada internetten alınan 3670 çiçekten oluşan veri seti kullanılarak sınıflandırma çalışması yapılmıştır. Son dönemde görüntü sınıflandırma çalışmalarında derin öğrenme yöntemleri kullanılarak oldukça başarılı sonuçlara ulaşılmaktadır. Bu çalışma derin öğrenme modellerinden ön eğitimli evrişimsel sinir ağları (ESA) AlexNet, GoogLeNet, SqueezeNet, ShuffleNet ve Resnet-18 ile sınıflandırma çalışması yapılarak performansları karşılaştırmalı olarak irdelenmiştir. Karşılaştırma neticesinde en başarılı sonuca %97.26 doğruluk oranına sahip olan GoogLeNet ile ulaşılmıştır. Diğer modeller için elde edilen başarı oranları sırasıyla ShuffleNet, SqueezeNet, ResNet-18 ve AlexNet için %97.23, %92.84, %91.42 %89.05’tir. Çalışmada GoogLeNet modeli bu çalışmadaki modellerin yanı sıra aynı veri seti ile yapılan diğer alışmalar içinde en yüksek başarıya ulaşan model olmuştur.

Anahtar Kelimeler

• Derin Öğrenme
• Sınıflandırma
• Evrişimsel Sinir Ağları
• çiçek
• GoogLeNet

Performance Comparison of Pre-Trained Convolutional Neural Networks in Flower Image Classification

Abstract

Flower classification is important for many fields from botonics to ecological studies. The difficulties encountered in classification studies are that flower images are not clearly evident, leaves and branches obscure the image, and there are many flowers with similar characteristics. In the study, classification study was carried out using the data set consisting of 3670 flowers taken from the internet. Recently, very successful results have been achieved by using deep learning methods in image classification studies. In this study, the performances of the deep learning models were examined comparatively by making a classification study with the pre-trained convolutional neural networks (CNN) AlexNet, GoogLeNet, SqueezeNet, ShuffleNet and Resnet-18. As a result of the comparison, the most successful result was obtained with GoogLeNet, which has an accuracy rate of 97.26%. The accuracy rate was calculated as 97.23%, 92.84%, 91.42% 89.05% for ShuffleNet, SqueezeNet, ResNet-18 and AlexNet, respectively. In addition to the models in this study, the GoogLeNet model in the study was the model that achieved the highest success among other studies conducted with the same data set.

Keywords

• Deep Learning
• Classification
• Convolutional Neural Networks
• flower
• GoogLeNet

Kaynakça

1. Acikgoz, H. (2022). A novel approach based on integration of convolutional neural networks and deep feature selection or short-term solar radiation forecasting. Applied Energy, (305). doi:https://doi.org/10.1016/j.apenergy.2021.117912.
2. Atik, I. (n.d.). COVID-19 Case Forecast with Deep Learning BiLSTM Approach: The Turkey Case. International Journal of Mechanical Engineering, 7(1), 6307–6314.
3. Cho, S.-Y., & Lim, P.-T. (2006). A novel Virus Infection Clustering for Flower Images Identification. In 18th International Conference on Pattern Recognition (ICPR’06) (Vol. 2, pp. 1038–1041). doi:10.1109/ICPR.2006.144
4. Coşkun, U. A., & Demi̇rhan, A. (2022). Farklı Çiçek Türlerini Derin Öğrenme Yöntemi İle Tanıma. Dokuz Eylül Üniversitesi Mühendislik Fakültesi Fen ve Mühendislik Dergisi, 24(70), 55–64. doi:10.21205/deufmd.2022247007
5. Das, M., Manmatha, R., & Riseman, E. M. (1999). Indexing flower patent images using domain knowledge. IEEE Intelligent Systems and Their Applications, 14(5), 24–33. doi:10.1109/5254.796084
6. Demir, F., Abdullah, D. A., & Sengur, A. (2020). A New Deep CNN Model for Environmental Sound Classification. IEEE Access, (8), 66529–66537.
7. Guo, B., Hu, J., Wu, W., Peng, Q., & Wu, F. (2019). The Tabu_Genetic Algorithm: A Novel Method for Hyper-Parameter Optimization of Learning Algorithms. Electronics, 8(5). doi:10.3390/electronics8050579
8. Guru, D., Kumar, Y. H., & Shantharamu, M. (2010). Texture Features and KNN in Classification of Flower Images. International Journal of Computer Applications,Special Issue on RTIPPR, 1, 21–29.

9. Hiary, H., Saadeh, H., Saadeh, M. K., & Yaqub, M. (2018). Flower classification using deep convolutional neural networks. IET Comput. Vis., 12, 855–862.
10. Kaggle. (2021, December 6). Kaggle. Kaggle data set. dataset. Retrieved from https://www.kaggle.com/datasets
11. Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). Imagenet classification with deep convolutional neural networks. Advances in Neural Information Processing Systems, 25, 1097–1105.
12. Liu, Y., Tang, F., Zhou, D., Meng, Y., & Dong, W. (2016). Flower classification via convolutional neural network. In 2016 IEEE International Conference on Functional-Structural Plant Growth Modeling, Simulation, Visualization and Applications (FSPMA) (pp. 110–116). doi:10.1109/FSPMA.2016.7818296
13. Luus, F. P. S., Khan, N., & Akhalwaya, I. (2019). Active Learning with TensorBoard Projector. CoRR, abs/1901.00675. Retrieved from http://arxiv.org/abs/1901.00675
14. Narayanan, B. N., Ali, R., & Hardie, R. C. (2019). Performance analysis of machine learning and deep learning architectures for malaria detection on cell images (Vol. 11139, p. 111390W). Presented at the Applications of Machine Learning, International Society for Optics and Photonics.
15. Nilsback, M.-E., & Zisserman, A. (2006). A Visual Vocabulary for Flower Classification. In 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’06) (Vol. 2, pp. 1447–1454). doi:10.1109/CVPR.2006.42
16. Nilsback, M.-E., & Zisserman, A. (2008). Automated Flower Classification over a Large Number of Classes. In 2008 Sixth Indian Conference on Computer Vision, Graphics Image Processing (pp. 722–729). doi:10.1109/ICVGIP.2008.47
17. Nilsback, M.-E., & Zisserman, A. (2010). Delving Deeper into the Whorl of Flower Segmentation. Image Vision Comput., 28(6), 1049–1062. doi:10.1016/j.imavis.2009.10.001
18. Sarıgül, M., Ozyildirim, B. M., & Avci, M. (2019). Differential convolutional neural network. Neural Networks, 116, 279–287. doi:https://doi.org/10.1016/j.neunet.2019.04.025
19. Shailendrakumar, M. M., Sachin, R. G., & Dattatraya, S. B. (2011). On Scale Invariance Texture Image Retrieval using Fuzzy Logic and Wavelet Co-occurrence based Features. International Journal of Computer Applications, 18, 10–17.
20. Szegedy, C., Ioffe, S., Vanhoucke, V., & Alemi, A. (2017). Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning. Proceedings of the AAAI Conference on Artificial Intelligence, 31(1). Retrieved from https://ojs.aaai.org/index.php/AAAI/article/view/11231
21. Toğaçar, M., Ergen, B., & Cömert, Z. (2019). A Deep Feature Learning Model for Pneumonia Detection Applying a Combination of mRMR Feature Selection and Machine Learning Models. IRBM. doi:10.1016/j.irbm.2019.10.006
22. Tseng, V. S., Wang, M.-H., & Su, J.-H. (2005). A New Method for Image Classification by Using Multilevel Association Rules. In 21st International Conference on Data Engineering Workshops (ICDEW’05) (pp. 1180–1180). doi:10.1109/ICDE.2005.164
23. Ucar, F., & Korkmaz, D. (2020). COVIDiagnosis-Net: Deep Bayes-SqueezeNet based diagnosis of the coronavirus disease 2019 (COVID-19) from X-ray images. Medical Hypotheses, 140, 109761–109761. doi:10.1016/j.mehy.2020.109761
24. Wu, Y., Qin, X., Pan, Y., & Yuan, C. (2018). Convolution Neural Network based Transfer Learning for Classification of Flowers. 2018 IEEE 3rd International Conference on Signal and Image Processing (ICSIP), 562–566.
25. Xia, X., Xu, C., & Nan, B. (2017). Inception-v3 for flower classification.
26. Zhang, X., Zhou, X., Lin, M., & Sun, J. (2017). ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices.