Advertisement

Improved Performance of CDL Algorithm Using DDELM-AE and AK-SVD

  • Xiulan YuEmail author
  • Junwei Mao
  • Chenquan Gan
  • Zufan Zhang
Conference paper
Part of the Lecture Notes in Electrical Engineering book series (LNEE, volume 516)

Abstract

Due to the poor robustness and high complexity of the concentrated dictionary learning (CDL) algorithm, this paper addresses these issues using denoising deep extreme learning machine based on autoencoder (DDELM-AE) and approximate k singular value decomposition (AK-SVD). Different from the CDL algorithm, on input, DDELM-AE is added for enhancing denoising ability and AK-SVD replaces K-SVD for improving running speed. Additionally, experimental results show that the improved algorithm is more efficient than the original CDL algorithm in terms of running time, denoising ability, and stability.

Keywords

Signal compression CDL Deep learning DDELM-AE AK-SVD 

Notes

Acknowledgments

This work is supported by Natural Science Foundation of China (Grant No. 61702066), Scientific and Technological Research Program of Chongqing Municipal Education Commission (Grant No. KJ1704080), and Chongqing Research Program of Basic Research and Frontier Technology (Grant No. cstc2017jcyjAX0256).

References

  1. 1.
    Aharon M, Elad M, Bruckstein A. K-SVD: design of dictionaries for sparse representation. In: Proceedings of the workshop on signal processing with adaptive sparse structured representations (SPARS05); 2005. p. 9–12.Google Scholar
  2. 2.
    Aharon M, Elad M, Bruckstein A. K-SVD: an algorithm for designing overcomplete dictionaries for sparse representation. IEEE Trans Image Process. 2006;54(11):4311–22.CrossRefGoogle Scholar
  3. 3.
    Wei F, Shutao L, Leyuan F, Benediktsson JA. Contextual online dictionary learning for hyperspectral image classification. IEEE Trans Geosci Remote Sens. 2017;1–12.Google Scholar
  4. 4.
    Gkioulekas IA, Zickler T. Dimensionality reduction using the sparse linear model. In: Proceedings of the 2011 advances in neural information processing systems; 2011. p. 271–9.Google Scholar
  5. 5.
    Calderbank R, Jafarpour S, Schapire R. Compressed learning: universal sparse dimensionality reduction and learning in the measurement domain. Princeton University, USA; 2009. Technical Report.Google Scholar
  6. 6.
    Donoho DL. Compressed sensing. IEEE Trans Inf Theor. 2006;52(4):1289–306.MathSciNetCrossRefGoogle Scholar
  7. 7.
    SiLong Z, YuanXiang L, Xian W, XiShuai P. Nonlinear dimensionality reduction based on dictionary learning. In: ACTA AUTOMATICA SINICA; 2016. p. 1065–76.Google Scholar
  8. 8.
    Xiangyi C, Huaping L, Xinying X, Fuchun S. Denoising deep extreme learning machine for sparse representation. Memetic Comp. 2017;9(3):199–212.CrossRefGoogle Scholar
  9. 9.
    Shuting C, Shaojia W, Binling L, Daolin H, Simin Y, Shuqiong X. A dictionary-learning algorithm based on method of optimal directions and approximate K-SVD. In: 2016 35th Chinese control conference (CCC); 2016. p. 6957–61.Google Scholar
  10. 10.
    Zeyde R, Elad M, Protter M. On single image scale-upusingsparse-representations. In: Proceedings of the 7th international conference on curves and surfaces; 2012. p. 711–30.Google Scholar
  11. 11.
    Jianchao Y, Wright J, Thoams H, Yi M. Image super resolution via sparse representation. IEEE Trans Image Process. 2010;19(11):2861–73.MathSciNetCrossRefGoogle Scholar
  12. 12.
    Vincent P, Larochelle H, Lajoie I, Bengio Y. Stacked denoising autoencoders: learning useful representations in a deep network with a local denoising criterion. J Mach Learn Res. 2010;11(6):3371–408.MathSciNetzbMATHGoogle Scholar

Copyright information

© Springer Nature Singapore Pte Ltd. 2020

Authors and Affiliations

  • Xiulan Yu
    • 1
    Email author
  • Junwei Mao
    • 1
  • Chenquan Gan
    • 1
  • Zufan Zhang
    • 1
  1. 1.Chongqing Key Labs of Mobile Communications TechnologyChongqing University of Posts and TelecommunicationsChongqingChina

Personalised recommendations