Abstract
Gait information can be collected by a long-distance camera. But the relative angle between the subject and the camera changes, resulting in a cross-view gait recognition problem. This paper proposes a view transformation model method based on feature separation generate adversarial networks. Based on the GAN model, this method separates the features of the input data as an additional discriminant basis. On the premise of building a single model, it can convert image to any angle as needed. In order to make the images generated by GAN more realistic, the proposed method separates view and dress information from the identity data and encodes them. The discriminator is also optimized by adding the conditional codes as an additional basis, so that the generator can generate the corresponding image more realistically based on the encoded information image. In addition, the proposed method also adds a constraint to increase the inter-class variation of subjects and reduce their intra-class distance. Thus, the synthesized image retains more feature information of original subject. The proposed method achieves a great generating effect and improves the performance of cross-view gait recognition.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Kusakunniran, W., Wu, Q., Li, H., Zhang, J.: Multiple views gait recognition using view transformation model based on optimized gait energy image. In: ICCV Workshops, pp. 1058–1064 (2009)
Makihara, Y., Suzuki, A., Muramatsu, D., Li, X., Yagi, Y.: Joint intensity and spatial metric learning for robust gait recognition. In: CVPR, pp. 5705–5715 (2017)
Singh, S., Biswas, K.: Biometric gait recognition with carrying and clothing variants. In: ICPRMI, pp. 446–451 (2009)
Johnson, A.Y., Bobick, A.F.: A multi-view method for gait recognition using static body parameters. In: 3rd International Conference on Audio and Video Based Biometric Person Authentication, pp. 301–311 (2001)
Bashir, K., Xiang, T., Gong, S.: Cross-view gait recognition using correlation strength. In: BMVC (2010)
Hu, M., Wang, Y., Zhang, Z., Little, J.J., Huang, D.: View-invariant discriminative projection for multi-view gait-based human identification. IEEE TIFS 8(12), 2034–2045 (2013)
Kusakunniran, W., Wu, Q., Li, H., Zhang, J.: Recognizing gaits across views through correlated motion co-clustering. IEEE Trans. Image Process. 23(2), 696–709 (2014)
Makihara, Y., Sagawa, R., Mukaigawa, Y., Echigo, T., Yagi, Y.: Gait recognition using a view transformation model in the frequency domain. In: Leonardis, A., Bischof, H., Pinz, A. (eds.) ECCV 2006. LNCS, vol. 3953, pp. 151–163. Springer, Heidelberg (2006). https://doi.org/10.1007/11744078_12
Goodfellow, I., et al.: Generative adversarial nets. In: NIPS (2014)
Chen, X., Duan, Y., Houthooft, R., Schulman, J., Sutskever, I., Abbeel, P.: InfoGAN: interpretable representation learning by information maximizing generative adversarial nets. In: NIPS (2016)
Tran, L., Yin, X., Liu, X.: Disentangled representation learning GAN for pose-invariant face recognition. In: CVPR, pp. 1283–1292 (2017)
Feng, Y., Li, Y., Luo, J.: Learning effective gait features using LSTM. In: International Conference on Pattern Recognition, pp. 325–330 (2017)
Yu, S., Chen, H., Wang, Q., Shen, L., Huang, Y.: Invariant feature extraction for gait recognition using only one uniform model. Neurocomputing 239, 81–93 (2017)
Yu, S., Chen, H., Garcia Reyes, E.B., Poh, N.: Gaitgan: Invariant gait feature extraction using generative adversarial networks. In: CVPR Workshops, pp. 532–539 (2017)
Han, J., Bhanu, B.: Individual recognition using gait energy image. IEEE TPAMI 28(2), 316–322 (2006)
Yu, S., Tan, D., Tan, T.: A framework for evaluating the effect of view angle, clothing and carrying condition on gait recognition. In: ICPR, pp. 441–444 (2006)
Yoo, D., Kim, N., Park, S., Paek, A.S., Kweon, I.S.: Pixel-level domain transfer. arXiv:CoRR 1603, 07442 (2016)
Deng, W., Zheng, L., Kang, G., Yang, Y., Ye, Q., Jiao, J.: Imageimage domain adaptation with preserved self-similarity and domaindissimilarity for person reidentification. In: CVPR, pp. 994–1003 (2018)
Kingma, D.P., Ba, J.: Adam: a method for stochastic optimization. arXiv preprint arXiv:1412.6980 (2014)
Yu, S., et al.: Gaitganv2: invariant gait feature extraction using generative adversarial networks. Pattern Recogn. 87, 179–189 (2019)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2021 Springer Nature Switzerland AG
About this paper
Cite this paper
Huang, C., Song, Y., Zhang, Y. (2021). Feature Separation GAN for Cross View Gait Recognition. In: Peng, Y., Hu, SM., Gabbouj, M., Zhou, K., Elad, M., Xu, K. (eds) Image and Graphics. ICIG 2021. Lecture Notes in Computer Science(), vol 12888. Springer, Cham. https://doi.org/10.1007/978-3-030-87355-4_6
Download citation
DOI: https://doi.org/10.1007/978-3-030-87355-4_6
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-87354-7
Online ISBN: 978-3-030-87355-4
eBook Packages: Computer ScienceComputer Science (R0)