Pedestrian Detection Fusing HOG Based on LE and Haar-Like Feature

  • Jin Huang
  • Bo Li
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 10956)


In order to change the low detecting speed and excess redundant information of the gradient histogram (HOG), this paper proposes the pedestrian detection fusing HOG based on Laplacian Eigenmaps (LE) dimensionality reduction and Haar-Like feature. By constructing a pedestrian detection sample set mixed with HOG and Haar-Like feature set, and using LE to reduce the HOG feature, then fuse Haar-Like feature extraction with HOG-LE feature. Finally, we use Adaboost cascade strong classifier to test INRIA Person sample library, the test results are obviously better than the single feature extraction method in terms of detection rate, and the text also compares the method with some other pedestrian detection methods. The pedestrian detection fusing HOG based on LE reduced dimension and Haar-Like features proposed in this paper improves detection rate effectively in real scenarios.


HOG Haar-Like LE Adaboost Pedestrian detection 



This work was supported by the National Natural Science Foundation of China (Grant no. 61273303, 61572381). The authors would like to thank all the editors and reviewers for their valuable comments and suggestions.


  1. 1.
    Nanda, H., Davis, L.: Probabilistic template based pedestrian detection in infrared videos. In: Intelligent Vehicle Symposium, vol. 1, pp. 15–20. IEEE (2008)Google Scholar
  2. 2.
    Ouyang, W., Wang, X.: Single-pedestrian detection aided by multi-pedestrian detection. In: IEEE Conference on Computer Vision and Pattern Recognition, pp. 3198–3205. IEEE Computer Society (2013)Google Scholar
  3. 3.
    Jiang, J., Xiong, H.: Fast pedestrian detection based on HOG-PCA and gentle AdaBoost. In: International Conference on Computer Science and Service System, pp. 1819–1822. IEEE (2012)Google Scholar
  4. 4.
    Ma, S., Bai, L.: A face detection algorithm based on Adaboost and new Haar-Like feature. In: IEEE International Conference on Software Engineering and Service Science, pp. 651–654. IEEE (2017)Google Scholar
  5. 5.
    Belkin, M., Niyogi, P.: Convergence of Laplacian eigenmaps. In: Proceedings of the Twentieth Conference on Advances in Neural Information Processing Systems, vol. 19, Vancouver, British Columbia, Canada, pp. 129–136. DBLP, December 2008Google Scholar
  6. 6.
    Taiana, M., Nascimento, J.C., Bernardino, A.: An improved labelling for the INRIA person data set for pedestrian detection. In: Sanches, J.M., Micó, L., Cardoso, J.S. (eds.) IbPRIA 2013. LNCS, vol. 7887, pp. 286–295. Springer, Heidelberg (2013). Scholar
  7. 7.
    Wei, Y., Tian, Q., Guo, T.: An improved pedestrian detection algorithm integrating haar-like features and HOG descriptors. In: Advances in Mechanical Engineering, p. 546206 (2013)CrossRefGoogle Scholar
  8. 8.
    Cheng, W.C., Jhan, D.M.: A cascade classifier using Adaboost algorithm and support vector machine for pedestrian detection. In: IEEE International Conference on Systems, Man, and Cybernetics, pp. 1430–1435. IEEE (2011)Google Scholar
  9. 9.
    Cao, J., Pang, Y., Li, X.: Pedestrian detection inspired by appearance constancy and shape symmetry. IEEE Trans. Image Process. 25(12), 5538–5551 (2016)MathSciNetCrossRefGoogle Scholar
  10. 10.
    Zou, C., Cai, D., Zhao, N., et al.: A pedestrian detection algorithm based on SVM-LeNet fusion. Comput. Eng. 43(5), 169–173 (2017). (in Chinese)Google Scholar
  11. 11.
    Luo, J., Qi, C., Zhang, J., et al.: Pedestrian detection based on multifractal spectrum. J. Tianjin Polytech. Univ. 36(2), 59–63 (2017). (in Chinese)Google Scholar

Copyright information

© Springer International Publishing AG, part of Springer Nature 2018

Authors and Affiliations

  1. 1.Hubei Province Key Laboratory of Intelligent Information Processing and Real-Time Industrial System, College of Computer Science and TechnologyWuhan University of Sciences and TechnologyWuhanChina

Personalised recommendations