Deep Learning with Limited Data: Advanced Classification Approaches Through Few-Shot Learning and Prototype Networks

Abstract

Classification problems in the fields of machine learning and artificial intelligence facilitate the extraction of meaningful information from data by assigning inputs to specific categories. Classification processes offer solutions for a wide range of areas, including health, agriculture, education, and sports. However, the classification process typically requires a large amount of labeled data. Accessing a large volume of labeled data is costly and time-consuming. The few-shot learning method has been utilized to address this issue, allowing models to learn new tasks with minimal examples. In this article, pre-trained deep network architectures have been fed into prototype networks, creating representative examples for each class. Thus, the category to which new data belongs is determined based on its similarity to the prototypes. Experimental studies have been conducted on the Food101 and Oxford-III Pet datasets, and the experimental results have been measured using four different evaluation metrics. The results have been presented and interpreted both in table form and graphically. In comparing classification accuracy, the metrics of Accuracy, F1_Score, Precision, and Recall were utilized. For the Oxford-III Pet dataset, ResNet18 demonstrated the best classification performance with metric values of 0.9986, 1, 1, and 1 for Accuracy, F1_Score, Precision, and Recall, respectively. In the case of the Food101 dataset, EfficientNetB0 achieved the highest classification performance, with values of 0.9320, 0.93, 0.94, and 0.93 for Accuracy, F1_Score, Precision, and Recall, respectively.

Keywords

• Classification
• few-shot learning
• transfer learning

Sınırlı Veri ile Derin Öğrenme: Birkaç Atışlı Öğrenme ve Prototip Ağlar Aracılığıyla Gelişmiş Sınıflandırma Yaklaşımları

Abstract

Sınıflandırma problemleri, makine öğrenimi ve yapay zekâ alanında, girdileri belirli kategorilere atayarak verilerden anlamlı bilgi çıkarılmasını sağlar. Sınıflandırma işlemleri; sağlık, tarım, eğitim ve spor gibi geniş bir alan için çözümler sunar. Ancak, sınıflandırma işlemi yapılırken genellikle büyük miktarda etiketli veriye ihtiyaç duyulur. Büyük miktarda etiketli veriye ulaşmak maliyetli ve zaman alıcıdır. Bu problemin çözebilmek için birkaç atışlı öğrenme yöntemi ile modelin çok sınırlı örneklerle yeni görevleri öğrenmesine olanak tanınmıştır. Bu makalede, önceden eğitilmiş derin ağ mimarileri prototip ağlara beslenmiş ve her sınıf için temsilci örnekler oluşturulmuştur. Böylece, yeni verilerin hangi kategoriye ait olduğu prototiplere olan benzerliğe göre belirlenmiştir. Deneysel çalışmalar, Food101 ve Oxford-III Pet veri setleri üzerinde denenmiş ve deneysel sonuçlar dört farklı değerlendirme metriği ile ölçülmüştür. Deneysel sonuçlar hem tablo olarak hem de grafiksel olarak gösterilmiş ve yorumlanmıştır. Sınıflandırma doğruluğunu karşılaştırmak için Doğruluk, F1_Skoru, Kesinlik ve Duyarlılık metrikleri kullanılmıştır. Oxford-III Pet veri seti için, ResNet18 mimarisi sırasıyla Doğruluk, F1_Skoru, Kesinlik ve Duyarlılık için 0.9986, 1, 1 ve 1 değerleriyle en iyi sınıflandırma performansını göstermiştir. Food101 veri seti için ise EfficientNetB0 mimarisi sırasıyla 0.9320, 0.93, 0.94 ve 0.93 değerleriyle Doğruluk, F1_Skoru, Kesinlik ve Duyarlılık açısından en yüksek sınıflandırma performansına ulaşmıştır.

Keywords

• Sınıflandırma
• az atışlı öğrenme
• transfer öğrenme

References

1. [1] Toptas, B., Hanbay, D. "The Separation of Glaucoma and Non-Glaucoma Fundus Images using EfficientNet-B0." Bitlis Eren Üniversitesi Fen Bilimleri Dergisi, vol. 11, No. 4, pp. 1084–1092, 2022.
2. [2] Gudla, R., Vollala, S., Srinivasa, K. G., Amin, R. "A novel approach for classification of Tor and non-Tor traffic using efficient feature selection methods." Expert Systems with Applications, Article ID 123544, 2024.
3. [3] Gündüz, A. F., Talu, M. F. "Atrial fibrillation classification and detection from ECG recordings." Biomedical Signal Processing and Control, vol. 82, Article ID 104531, 2023.
4. [4] Toptaş, B., Hanbay, D. "Retinal blood vessel segmentation using pixel-based feature vector." Biomedical Signal Processing and Control, vol. 70, Article ID 103053, 2021.
5. [5] Toptaş, M. Orman yangınlarının görüntü işleme yöntemleri ile tespit edilmesi ve sınıflandırılması (Yüksek Lisans tezi, İnönü Üniversitesi Fen Bilimleri Enstitüsü), 2018.
6. [6] Krizhevsky, A., Sutskever, I., Hinton, G. E. "Imagenet classification with deep convolutional neural networks." Advances in Neural Information Processing Systems, vol. 25, 2012. [7] Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., et al. "Going deeper with convolutions." Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 1–9, 2015.
7. [8] Simonyan, K., Zisserman, A. "Very deep convolutional networks for large-scale image recognition." arXiv preprint arXiv:1409.1556, 2014.
8. [9] He, K., Zhang, X., Ren, S., Sun, J. "Deep residual learning for image recognition." Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 770–778, 2016.

9. [10] Huang, G., Liu, Z., Van Der Maaten, L., Weinberger, K. Q. "Densely connected convolutional networks." Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 4700–4708, 2017.
10. [11] Fei-Fei, L., Fergus, R., Perona, P. "One-shot learning of object categories." IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 28, No. 4, pp. 594–611, 2006.
11. [12] Paeedeh, N., Pratama, M., Ma’sum, M. A., Mayer, W., Cao, Z., Kowlczyk, R. "Cross-domain few-shot learning via adaptive transformer networks." Knowledge-Based Systems, Article ID 111458, 2024.
12. [13] Zhao, P., Wang, L., Zhao, X., Liu, H., Ji, X. "Few-shot learning based on prototype rectification with a self-attention mechanism." Expert Systems with Applications, Article ID 123586, 2024.
13. [14] Snell, J., Swersky, K., Zemel, R. "Prototypical networks for few-shot learning." Advances in Neural Information Processing Systems, pp. 4077–4087, 2017.
14. [15] Sung, F., Yang, Y., Zhang, L., Xiang, T., Torr, P. H., Hospedales, T. M. "Learning to compare: Relation network for few-shot learning." Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 1199–1208, 2018.
15. [16] Wang, Y., Chao, W. L., Weinberger, K. Q., Van Der Maaten, L. "Simpleshot: Revisiting nearest-neighbor classification for few-shot learning." arXiv preprint arXiv:1911.04623, 2019.
16. [17] Gülcü, A., Alkan, M. "Az Örnekle Öğrenme Problemleri için MAML ve ProtoNet Algoritmalarının İncelenmesi." Avrupa Bilim ve Teknoloji Dergisi, No. 21, pp. 113–121, 2021.
17. [18] Işık, G. "Improving Plant Disease Recognition Through Gradient-Based Few-shot Learning with Attention Mechanisms." Journal of the Institute of Science and Technology, vol. 13, No. 3, pp. 1482–1495, 2023.
18. [19] Argüeso, D., Picon, A., Irusta, U., Medela, A., San-Emeterio, M. G., Bereciartua, A., Alvarez-Gila, A. "Few-Shot Learning approach for plant disease classification using images taken in the field." Computers and Electronics in Agriculture, vol. 175, Article ID 105542, 2020.
19. [20] Wang, B., Wang, D. "Plant leaves classification: A few-shot learning method based on siamese network." IEEE Access, vol. 7, pp. 151754–151763, 2019.
20. [21] Frikha, A., et al. "Few-shot one-class classification via meta-learning." Proceedings of the AAAI Conference on Artificial Intelligence, vol. 35, No. 8, 2021.
21. [22] Chen, D., Chen, Y., Li, Y., Mao, F., He, Y., Xue, H. "Self-supervised learning for few-shot image classification." ICASSP 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 1745–1749, IEEE, June 2021.
22. [23] Krenzer, A., Heil, S., Fitting, D., Matti, S., Zoller, W. G., Hann, A., Puppe, F. "Automated classification of polyps using deep learning architectures and few-shot learning." BMC Medical Imaging, vol. 23, No. 1, Article ID 59, 2023.
23. [24] Liu, B., Yu, X., Yu, A., Zhang, P., Wan, G., Wang, R. "Deep few-shot learning for hyperspectral image classification." IEEE Transactions on Geoscience and Remote Sensing, vol. 57, No. 4, pp. 2290–2304, 2018.
24. [25] Kang, D., Cho, M. "Integrative few-shot learning for classification and segmentation." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 9979–9990, 2022.
25. [26] Hu, Y., Gripon, V., Pateux, S. "Leveraging the feature distribution in transfer-based few-shot learning." International Conference on Artificial Neural Networks, pp. 487–499, Springer, Cham, September 2021.
26. [27] Kim, J., Kim, T., Kim, S., Yoo, C. D. "Edge-labeling graph neural network for few-shot learning." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 11–20, 2019.
27. [28] Parkhi, O. M., Vedaldi, A., Zisserman, A., Jawahar, C. V. "Cats and dogs." 2012 IEEE Conference on Computer Vision and Pattern Recognition, pp. 3498–3505, https://doi.org/10.1109/CVPR.2012.6248092, 2012.
28. [29] Bossard, L., Guillaumin, M., Van Gool, L. "Food-101 – Mining discriminative components with random forests." European Conference on Computer Vision (ECCV), pp. 446–461, https://doi.org/10.1007/978-3-319-10599-4_29, 2014.
29. [30] Gao, T., Han, X., Zhu, H., Liu, Z., Li, P., Sun, M., Zhou, J. "FewRel 2.0: Towards more challenging few-shot relation classification." arXiv preprint arXiv:1910.07124, 2019.
30. [31] Singh, P., Mazumder, P. "Dual class representation learning for few-shot image classification." Knowledge-Based Systems, vol. 238, Article ID 107840, 2022.
31. [32] Ren, H., Cai, Y., Lau, R. Y., Leung, H. F., Li, Q. "Granularity-aware area prototypical network with bimargin loss for few-shot relation classification." IEEE Transactions on Knowledge and Data Engineering, vol. 35, No. 5, pp. 4852–4866, 2022.
32. [33] Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., Fei-Fei, L. "ImageNet: A large-scale hierarchical image database." 2009 IEEE Conference on Computer Vision and Pattern Recognition, pp. 248–255, https://doi.org/10.1109/CVPR.2009.5206848, 2009.
33. [34] Tan, M., Le, Q. "EfficientNet: Rethinking model scaling for convolutional neural networks." International Conference on Machine Learning, pp. 6105–6114, PMLR, May 2019.
34. [35] Howard, A., Sandler, M., Chu, G., Chen, L. C., Chen, B., Tan, M., et al. "Searching for MobileNetV3." Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 1314–1324, 2019.