Page 41 - First special issue on The impact of Artificial Intelligence on communication networks and services

,78 -2851$/ ,&7 'LVFRYHULHV 9RO 0DUFK

[32] S. Ioffe and C. Szegedy, “Batch normalization: Accel- [44] M. Everingham, L. Van Gool, C. K. Williams, J. Winn,
erating deep network training by reducing internal co- and A. Zisserman, “The pascal visual object classes
variate shift,” in International Conference on Machine (VOC) challenge,” International journal of computer
Learning, 2015, pp. 448–456. vision, vol. 88, no. 2, pp. 303–338, 2010.
[33] C. Szegedy, V. Vanhoucke, S. Ioffe, J. Shlens, and [45] A. Geiger, P. Lenz, and R. Urtasun, “Are we ready for
Z. Wojna, “Rethinking the inception architecture for autonomous driving? the kitti vision benchmark suite,”
computer vision,” in Proceedings of the IEEE Con- in Computer Vision and Pattern Recognition (CVPR),
ference on Computer Vision and Pattern Recognition, 2012 IEEE Conference on. IEEE, 2012, pp. 3354–
2016, pp. 2818–2826. 3361.

[34] C. Szegedy, S. Ioffe, V. Vanhoucke, and A. A. Alemi, [46] Z. Cai, Q. Fan, R. S. Feris, and N. Vasconcelos, “A
“Inception-v4, inception-resnet and the impact of resid- unified multi-scale deep convolutional neural network
ual connections on learning.” in AAAI, 2017, pp. 4278– for fast object detection,” in European Conference on
4284. Computer Vision. Springer, 2016, pp. 354–370.

[35] K. Simonyan and A. Zisserman, “Very deep convo- [47] Y. Xiang, W. Choi, Y. Lin, and S. Savarese,
lutional networks for large-scale image recognition,” “Subcategory-aware convolutional neural networks for
arXiv preprint arXiv:1409.1556, 2014. object proposals and detection,” in Applications of
Computer Vision (WACV), 2017 IEEE Winter Confer-
[36] K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual ence on. IEEE, 2017, pp. 924–933.
learning for image recognition,” in Proceedings of the
[48] F. Yang, W. Choi, and Y. Lin, “Exploit all the layers:
IEEE conference on computer vision and pattern recog-
Fast and accurate cnn object detector with scale depen-
nition, 2016, pp. 770–778.
dent pooling and cascaded rejection classifiers,” in Pro-
[37] O. Russakovsky, J. Deng, H. Su, J. Krause, S. Satheesh, ceedings of the IEEE Conference on Computer Vision
S. Ma, Z. Huang, A. Karpathy, A. Khosla, M. Bernstein and Pattern Recognition, 2016, pp. 2129–2137.
et al., “Imagenet large scale visual recognition chal-
[49] X. Chen, K. Kundu, Y. Zhu, A. G. Berneshawi, H. Ma,
lenge,” International Journal of Computer Vision, vol.
S. Fidler, and R. Urtasun, “3d object proposals for ac-
115, no. 3, pp. 211–252, 2015.
curate object class detection,” in Advances in Neural
[38] A. Sharif Razavian, H. Azizpour, J. Sullivan, and Information Processing Systems, 2015, pp. 424–432.
S. Carlsson, “CNN features off-the-shelf: an astound- [50] X. Chen, K. Kundu, Z. Zhang, H. Ma, S. Fidler, and
ing baseline for recognition,” in Proceedings of the R. Urtasun, “Monocular 3d object detection for au-
IEEE conference on computer vision and pattern recog- tonomous driving,” in Proceedings of the IEEE Con-
nition workshops, 2014, pp. 806–813.
ference on Computer Vision and Pattern Recognition,
[39] R. Girshick, J. Donahue, T. Darrell, and J. Malik, “Rich 2016, pp. 2147–2156.
feature hierarchies for accurate object detection and se-
[51] NICS Lab of Tsinghua University, “Neural
mantic segmentation,” in Proceedings of the IEEE con-
network accelerator inference,” 2017. [Online].
ference on computer vision and pattern recognition,
Available: https://nicsefc.ee.tsinghua.edu.cn/projects/
2014, pp. 580–587.
neural-network-accelerator/
[40] K. He, X. Zhang, S. Ren, and J. Sun, “Spatial pyra- [52] NVIDIA, “NVIDIA Tesla V100,” 2017. [Online].
mid pooling in deep convolutional networks for visual
Available: https://www.nvidia.com/en-us/data-center/
recognition,” in European Conference on Computer Vi-
tesla-v100/
sion. Springer, 2014, pp. 346–361.
[53] ——, “NVIDIA Jetson - the embedded plat-
[41] R. Girshick, “Fast R-CNN,” in Proceedings of the IEEE form for autonomous everything,” 2017. [Online].
international conference on computer vision, 2015, pp. Available: http://www.nvidia.com/object/embedded-
1440–1448. systems-dev-kits-modules.html

[42] S. Ren, K. He, R. Girshick, and J. Sun, “Faster R-CNN: [54] I. Kuon and J. Rose, “Measuring the gap between fp-
Towards real-time object detection with region proposal gas and asics,” IEEE Transactions on computer-aided
networks,” in Advances in neural information process- design of integrated circuits and systems, vol. 26, no. 2,
ing systems, 2015, pp. 91–99. pp. 203–215, 2007.

[43] J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, [55] T. Chen, Z. Du, N. Sun, J. Wang, C. Wu,
“You only look once: Unified, real-time object detec- Y. Chen, and O. Temam, “Diannao: A small-footprint
tion,” in Proceedings of the IEEE Conference on Com- high-throughput accelerator for ubiquitous machine-
puter Vision and Pattern Recognition, 2016, pp. 779– learning,” in ACM Sigplan Notices, vol. 49, no. 4.
788. ACM, 2014, pp. 269–284.

‹ ,QWHUQDWLRQDO 7HOHFRPPXQLFDWLRQ 8QLRQ