Abstract
Few-shot learning in computer vision is a very challenging task. Many approaches have been proposed to tackle the few-shot learning problem. Meta learning, a method of learning to learn, is introduced into few-shot learning problem and has achieved pretty good results. But there is still a very big gap between the machine and our human in the few-shot learning tasks. We think it’s because the existing methods do not make full use of global knowledge (similar to the priori knowledge of human understanding of images) thus lacking a world view of the task. In other words, they focus too much on local information and neglect the whole task. In this paper, we rethink about the few-shot learning problem, and propose that we should make full use of global knowledge. Seen data set is used to obtain a embedding function between images and feature vectors, the images are embedded onto a hypersphere in the manner of cosine embedding. By taking this embedding function as global knowledge, we train a classifier to classify the corresponding embedded vector of images. The experiment proved that our approach significantly outperforms both baseline models and previous state-of-the-art methods. It surpasses most existing methods in terms of flexibility, simplicity and accuracy. Codes are available at https://github.com/SongyiGao/OICEFFSL.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Chen, W.Y., Liu, Y.C., Kira, Z., Wang, Y.C.F., Huang, J.B.: A closer look at few-shot classification. arXiv preprint arXiv:1904.04232 (2019)
Cheny, Z., Fuy, Y., Zhang, Y., Jiang, Y.G., Xue, X., Sigal, L.: Multi-level semantic feature augmentation for one-shot learning. IEEE Trans. Image Process. 28(9), 4594–4605 (2019)
Finn, C., Abbeel, P., Levine, S.: Model-agnostic meta-learning for fast adaptation of deep networks. In: Proceedings of the 34th International Conference on Machine Learning, vol. 70, pp. 1126–1135. JMLR. org (2017)
Gidaris, S., Komodakis, N.: Dynamic few-shot visual learning without forgetting. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 4367–4375 (2018)
He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 770–778 (2016)
Ioffe, S., Szegedy, C.: Batch normalization: accelerating deep network training by reducing internal covariate shift. arXiv preprint arXiv:1502.03167 (2015)
LeCun, Y., Bengio, Y., Hinton, G.: Deep learning. Nature 521(7553), 436 (2015)
Li, X., Wang, W.: Learning discriminative features via weights-biased softmax loss. arXiv preprint arXiv:1904.11138 (2019)
Long, J., Shelhamer, E., Darrell, T.: Fully convolutional networks for semantic segmentation. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 3431–3440 (2015)
Maaten, L.v.d., Hinton, G.: Visualizing data using T-SNE. J. Mach. Learn. Res. 9(Nov), 2579–2605 (2008)
Mishra, N., Rohaninejad, M., Chen, X., Abbeel, P.: A simple neural attentive meta-learner. arXiv preprint arXiv:1707.03141 (2017)
Nichol, A., Achiam, J., Schulman, J.: On first-order meta-learning algorithms. arXiv preprint arXiv:1803.02999 (2018)
Qin, Y., et al.: Rethink and redesign meta learning. arXiv preprint arXiv:1812.04955 (2018)
Ronneberger, O., Fischer, P., Brox, T.: U-Net: convolutional networks for biomedical image segmentation. In: Navab, N., Hornegger, J., Wells, W.M., Frangi, A.F. (eds.) MICCAI 2015. LNCS, vol. 9351, pp. 234–241. Springer, Cham (2015). https://doi.org/10.1007/978-3-319-24574-4_28
Russakovsky, O., et al.: Imagenet large scale visual recognition challenge. Int. J. Comput. Vis. 115(3), 211–252 (2015)
Salimans, T., Kingma, D.P.: Weight normalization: a simple reparameterization to accelerate training of deep neural networks. In: Advances in Neural Information Processing Systems, pp. 901–909 (2016)
Snell, J., Swersky, K., Zemel, R.: Prototypical networks for few-shot learning. In: Advances in Neural Information Processing Systems, pp. 4077–4087 (2017)
Sun, Q., Liu, Y., Chua, T.S., Schiele, B.: Meta-transfer learning for few-shot learning. arXiv preprint arXiv:1812.02391 (2018)
Sung, F., Yang, Y., Zhang, L., Xiang, T., Torr, P.H., Hospedales, T.M.: Learning to compare: relation network for few-shot learning. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 1199–1208 (2018)
Vinyals, O., Blundell, C., Lillicrap, T., Wierstra, D., et al.: Matching networks for one shot learning. In: Advances in Neural Information Processing Systems, pp. 3630–3638 (2016)
Wah, C., Branson, S., Welinder, P., Perona, P., Belongie, S.: The Caltech-UCSD Birds-200-2011 dataset (2011)
Zhang, H., Cisse, M., Dauphin, Y.N., Lopez-Paz, D.: mixup: beyond empirical risk minimization. arXiv preprint arXiv:1710.09412 (2017)
Author information
Authors and Affiliations
Corresponding authors
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2019 Springer Nature Switzerland AG
About this paper
Cite this paper
Gao, S., Shen, W., Liu, Z., Zhu, A., Yu, Y. (2019). Only Image Cosine Embedding for Few-Shot Learning. In: Gedeon, T., Wong, K., Lee, M. (eds) Neural Information Processing. ICONIP 2019. Lecture Notes in Computer Science(), vol 11954. Springer, Cham. https://doi.org/10.1007/978-3-030-36711-4_8
Download citation
DOI: https://doi.org/10.1007/978-3-030-36711-4_8
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-36710-7
Online ISBN: 978-3-030-36711-4
eBook Packages: Computer ScienceComputer Science (R0)