Pyramidal Combination of Separable Branches for Deep Short Connected Neural Networks

  • Yao Lu
  • Guangming LuEmail author
  • Rui Lin
  • Bing Ma
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 11257)


Recent works have shown that Convolutional Neural Networks (CNNs) with deeper structure and short connections have extremely good performance in image classification tasks. However, deep short connected neural networks have been proven that they are merely ensembles of relatively shallow networks. From this point, instead of traditional simple module stacked neural networks, we propose Pyramidal Combination of Separable Branches Neural Networks (PCSB-Nets), whose basic module is deeper, more delicate and flexible with much fewer parameters. The PCSB-Nets can fuse the caught features more sufficiently, disproportionately increase the efficiency of parameters and improve the model’s generalization and capacity abilities. Experiments have shown this novel architecture has improvement gains on benchmark CIFAR image classification datasets.


Deep learning CNNs PCSB-Nets 



The work is supported by the NSFC fund (61332011), Shenzhen Fundamental Research fund (JCYJ20170811155442454, GRCK2017042116121208), and Medical Biometrics Perception and Analysis Engineering Laboratory, Shenzhen, China.


  1. 1.
    Chollet, F.: Deep learning with separable convolutions. arXiv preprint arXiv:1610.02357 (2016)
  2. 2.
    Glorot, X., Bordes, A., Bengio, Y.: Deep sparse rectifier neural networks. J. Mach. Learn. Res. 15 (2011)Google Scholar
  3. 3.
    He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. In: The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 770–778 (2016)Google Scholar
  4. 4.
    Huang, G., Liu, Z., Weinberger, K.Q.: Densely connected convolutional networks (2016)Google Scholar
  5. 5.
    Huang, G., Sun, Y., Liu, Z., Sedra, D., Weinberger, K.Q.: Deep networks with stochastic depth. In: Leibe, B., Matas, J., Sebe, N., Welling, M. (eds.) ECCV 2016. LNCS, vol. 9908, pp. 646–661. Springer, Cham (2016). Scholar
  6. 6.
    Ioffe, S., Szegedy, C.: Batch normalization: accelerating deep network training by reducing internal covariate shift. Computer Science (2015)Google Scholar
  7. 7.
    Jia, Y., Shelhamer, E., Donahue, J., Karayev, S., Long, J.: Caffe: convolutional architecture for fast feature embedding. Eprint Arxiv pp. 675–678 (2014)Google Scholar
  8. 8.
    Krizhevsky, A., Sutskever, I., Hinton, G.E.: Imagenet classification with deep convolutional neural networks. Adv. Neural Inf. Process. Syst. 25(2), 2012 (2012)Google Scholar
  9. 9.
    Larsson, G., Maire, M., Shakhnarovich, G.: FractalNet: ultra-deep neural networks without residuals (2016)Google Scholar
  10. 10.
    Lecun, Y., Bottou, L., Bengio, Y., Haffner, P.: Gradient-based learning applied to document recognition. Proc. IEEE 86(11), 2278–2324 (1998)CrossRefGoogle Scholar
  11. 11.
    Sutskever, I., Martens, J., Dahl, G., Hinton, G.: On the importance of initialization and momentum in deep learning. In: International Conference on Machine Learning (2013)Google Scholar
  12. 12.
    Szegedy, C., Liu, W., Jia, Y., Sermanet, P.: Going deeper with convolutions. In: Computer Vision and Pattern Recognition, pp. 1–9 (2014)Google Scholar
  13. 13.
    Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., Wojna, Z.: Rethinking the inception architecture for computer vision. Computer Science (2015)Google Scholar
  14. 14.
    Zagoruyko, S., Komodakis, N.: Wide residual networks. arXiv preprint arXiv:1605.07146 (2016)

Copyright information

© Springer Nature Switzerland AG 2018

Authors and Affiliations

  1. 1.Harbin Institute of Technology (ShenZhen)ShenZhenChina

Personalised recommendations