A Semi-Supervised Active Learning Framework for Image Classification

Abstract:

In this paper, a novel image classification method, incorporating active learning and semi-supervised learning (SSL), is proposed. In this method, two classifiers are needed where one is trained by labeled data and some unlabeled data, while the other one is trained only by labeled data. Specifically, in each round, two classifiers iterate to select useful examples in contention for user query. Then we compute the label changing rate for every unlabeled example in each classifier. Those examples in which the label changing rate is zero and the label in the two classifiers is the same are selected to add into the training data of the first classifier. Our experimental results show that our method significantly reduced the need of labeled examples, while at the same time reducing classification error compared with widely used image classification algorithms.

• Periodical: Applied Mechanics and Materials (Volumes 556-562)
• Pages: 4765-4769
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.556-562.4765
• Online since: May 2014
• Authors: Han Yi Li*, Ming Yang, Nan Nan Kang, Lu Lu Yue
• Keywords: Active Learning, Image Classification, Semi-Supervised Learning (SSL)

References

[1] Zhang, Yihao, et al. Semi-supervised learning combining co-training with active learning., Expert Systems with Applications 41. 5 (2014): 2372-2378.

DOI: 10.1016/j.eswa.2013.09.035

[2] Settles, Burr. Active learning literature survey., University of Wisconsin, Madison 52 (2010): 55-66.

[3] Zhu, Xiaojin. Semi-supervised learning literature survey., Computer Science, University of Wisconsin-Madison 2 (2006): 3.

[4] Guillaumin, Matthieu, Jakob Verbeek, and Cordelia Schmid. Multimodal semi-supervised learning for image classification., Computer Vision and Pattern Recognition (CVPR), 2010 IEEE Conference on. IEEE, (2010).

DOI: 10.1109/cvpr.2010.5540120

[5] Joachims, Thorsten. Transductive inference for text classification using support vector machines., ICML. Vol. 99. (1999).

[6] Nigam, Kamal, et al. Text classification from labeled and unlabeled documents using EM., Machine learning 39. 2-3 (2000): 103-134.

[7] Zhou, Zhi-Hua, and Ming Li. Tri-training: Exploiting unlabeled data using three classifiers., Knowledge and Data Engineering, IEEE Transactions on17. 11 (2005): 1529-1541.

DOI: 10.1109/tkde.2005.186

[8] Blum, Avrim, and Tom Mitchell. Combining labeled and unlabeled data with co-training., Proceedings of the eleventh annual conference on Computational learning theory. ACM, (1998).

DOI: 10.1145/279943.279962

[9] Leng, Yan, Xinyan Xu, and Guanghui Qi. Combining active learning and semi-supervised learning to construct SVM classifier., Knowledge-Based Systems 44 (2013): 121-131.

DOI: 10.1016/j.knosys.2013.01.032

[10] Joshi, Ajay J., Fatih Porikli, and Nikolaos Papanikolopoulos. Multi-class active learning for image classification., Computer Vision and Pattern Recognition, 2009. CVPR 2009. IEEE Conference on. IEEE, (2009).

DOI: 10.1109/cvpr.2009.5206627