244
Y. Zhu et al.
channel. The infrared image derived from the cascaded visible and the target segmentation map has good performance. Depending on the effect, the outline of the target is
complete and the detail effect is effective. In the derivation of some scenes, the derived
image of GB is generally bright, but the target is still clearly identifiable and distinct
from the background. Since the details of GB generation are clearer, a slight inconsistency with the original image will cause a sharp drop in PSNR and SSIM, so the PSNR
and SSIM of the blurred image are higher, and these two indicators do not indicate the
generation effect very well.
References
1. Mescheder, L., Geiger, A., Nowozin, S.: Which training methods for GANs do actually
Converge? arXiv preprint arXiv:1801.04406 (2018)
2. Kniaz, V.V., Knyaz, V.A., Hlad˚ uvka, J., Kropatsch, W.G., Mizginov, V.: ThermalGAN: multimodal color-to-thermal image translation for person re-identification in multispectral dataset.
In: Leal-Taixé, L., Roth, S. (eds.) ECCV 2018. LNCS, vol. 11134, pp. 606–624. Springer,
Cham (2019). https://doi.org/10.1007/978-3-030-11024-6_46
3. Zheng, S., Jayasumana, S., Romeraparedes, B., et al.: Conditional random fields as recurrent
neural networks. In: International Conference on Computer Vision, vol. 2015, pp. 1529–1537
(2015)
4. Lin, M., Chen, Q., Yan, S., et al.: Network in network. In: International Conference on Learning
Representations (2014)
5. Arjovsky, M., Chintala, S., Bottou, L.: Wasserste in generative adversarial networks. In:
International Conference on Machine Learning, pp. 214–223 (2017)
6. Gulrajani, I., Ahmed, F., Arjovsky, M, et al.: Improved training of wasserstein gans. Adv.
Neural Inf. Process. Syst. 5767–5777 (2017)
7. Isola, P., Zhu, J., Zhou, T., et al.: Image-to-Image Translation with Conditional Adversarial
Networks. In: Computer Vision and Pattern Recognition, pp. 5967–5976 (2017)
Y. Zhu et al.
channel. The infrared image derived from the cascaded visible and the target segmentation map has good performance. Depending on the effect, the outline of the target is
complete and the detail effect is effective. In the derivation of some scenes, the derived
image of GB is generally bright, but the target is still clearly identifiable and distinct
from the background. Since the details of GB generation are clearer, a slight inconsistency with the original image will cause a sharp drop in PSNR and SSIM, so the PSNR
and SSIM of the blurred image are higher, and these two indicators do not indicate the
generation effect very well.
References
1. Mescheder, L., Geiger, A., Nowozin, S.: Which training methods for GANs do actually
Converge? arXiv preprint arXiv:1801.04406 (2018)
2. Kniaz, V.V., Knyaz, V.A., Hlad˚ uvka, J., Kropatsch, W.G., Mizginov, V.: ThermalGAN: multimodal color-to-thermal image translation for person re-identification in multispectral dataset.
In: Leal-Taixé, L., Roth, S. (eds.) ECCV 2018. LNCS, vol. 11134, pp. 606–624. Springer,
Cham (2019). https://doi.org/10.1007/978-3-030-11024-6_46
3. Zheng, S., Jayasumana, S., Romeraparedes, B., et al.: Conditional random fields as recurrent
neural networks. In: International Conference on Computer Vision, vol. 2015, pp. 1529–1537
(2015)
4. Lin, M., Chen, Q., Yan, S., et al.: Network in network. In: International Conference on Learning
Representations (2014)
5. Arjovsky, M., Chintala, S., Bottou, L.: Wasserste in generative adversarial networks. In:
International Conference on Machine Learning, pp. 214–223 (2017)
6. Gulrajani, I., Ahmed, F., Arjovsky, M, et al.: Improved training of wasserstein gans. Adv.
Neural Inf. Process. Syst. 5767–5777 (2017)
7. Isola, P., Zhu, J., Zhou, T., et al.: Image-to-Image Translation with Conditional Adversarial
Networks. In: Computer Vision and Pattern Recognition, pp. 5967–5976 (2017)