Deep Convolution Neural Network Recognition Algorithm Based on Maximum Scatter Difference Criterion

  • Kunlun LiEmail author
  • Xuefei Geng
  • Weiduan Li
Conference paper
Part of the Communications in Computer and Information Science book series (CCIS, volume 699)


Convolution neural network is a method that can extract features automatically of deep learning. It has a better recognition effect compared with a variety of face recognition algorithms. In view of the problem that the number of face training samples is reduced and the recognition performance is reduced too, the recognition algorithm based on maximum scatter difference criterion is proposed. The maximum scatter difference criterion is introduced to minimize the error when the gradient descent method is used to adjust the weight. And the within-class scatter of the sample should be the minimum and the between-class should be the maximum. Finally, the weights can be more close to the optimal value of the classification and the recognition rate of the system can be improved. A large number of experiments show the effectiveness of the algorithm.


Deep learning Convolution network Maximum scatter difference criterion Face recognition 



This work is supported by the National Science and Technology Support Program (2013BAK07B00), the Natural Science Foundation of Heibei Province of China under granted (F2013201170), the Educational Commission of Hebei Province of China (ZD2014008) and the National Natural Science Foundation of China (No. 61672205).


  1. 1.
    Gui, J., Sun, Z., Jia, W., et al.: Discriminant sparse neighborhood preserving embedding for face recognition. J. Pattern Recogn. 45, 2884–2897 (2012)CrossRefzbMATHGoogle Scholar
  2. 2.
    Geng, C., Jiang, X.D.: Fully automatic face recognition framework based on local and global features. J Mach. Vis. Appl. 19(3), 549–571 (2013)Google Scholar
  3. 3.
    Hinton, G.E., Osindero, S.: A fast learning algorithm for deep belief nets. J. Neural Comput. 18(7), 1527–1554 (2006)MathSciNetCrossRefzbMATHGoogle Scholar
  4. 4.
    Sun, Z., Xue, L., Xu, Y., et al.: Review on the research of deep learning. J. Comput. Appl. Res. 29(8), 2806–2810 (2012)Google Scholar
  5. 5.
    Lecun, Y.L., Bottou, L., Bengio, Y., et al.: Gradient-based learning applied to document recognition. J. Proc. IEEE 86(11), 2278–2324 (1998)CrossRefGoogle Scholar
  6. 6.
    Tang, P., Wang, H.: The depth of the parallel cross convolution neural network model. J. Chin. J. Image Graph. (2016)Google Scholar
  7. 7.
  8. 8.
    Zheng, N., Qi, L., Guan, L.: Generalized multiple maximum scatter difference feature extraction using QR decomposition. J. Vis. Commun. Image Represent. 25(6), 1460–1471 (2014)CrossRefGoogle Scholar

Copyright information

© Springer Nature Singapore Pte Ltd. 2017

Authors and Affiliations

  1. 1.College of Electronic and Information EngineeringHebei UniversityBaodingChina
  2. 2.College of Civil Engineering and ArchitectureHebei UniversityBaodingChina

Personalised recommendations