An eigendecomposition method based on deep learning and probabilistic graph model

Abstract

With the rapid development of computer, computer vision derived from computer vision has also made important progress in the field of image research. The extraction of image information is the most basic work in the field of image research. However, in the current environment, there is still a lack of effective methods to understand more complex image problems, such as image shape, material and illumination distribution in the environment. Eigenimage decomposition can be achieved by obtaining albedo eigenvalues and luminance eigenvalues. The color and illumination information of the image can be obtained more intuitively. Based on this, this paper proposes an intrinsic image decomposition method based on depth learning and probability graph model, in order to extract image information more accurately. Firstly, a deep convolution neural network is trained to decompose reflectivity image and shadow image. Then the conditional random field is used to optimize the reflectivity image and shadow image. The convolutional neural network designed in this paper obtains preliminary results through multi-scale architecture, deep supervision, step-by-step refinement of synthetic images and multi-stage training, which has been significantly improved compared with previous algorithms. Then the essential image and the corresponding gradient image are further optimized by conditional random field, and the eigenvalue image with richer details and clearer boundary can be obtained.

This is a preview of subscription content, log in to check access.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7

References

  1. Bell S, Bala K, Suavely N (2014) Intrinsic images in the wild. ACM Trans Graph (TOG) 33(4):159

    Article  Google Scholar 

  2. Bian Z, Tang P, Yan J (2019) Land-cover classification from multiple classifiers using decision fusion based on the probabilistic graphical model. Int J Remote Sens 40(12):1–17

    Article  Google Scholar 

  3. Dong J (2015) Research on fast and reliable template image matching technology. National University of Defense Science and Technology

  4. Du J (2017) Research on pixel-level multiscale medical image fusion method. Chongqing University of Posts and Telecommunications

  5. Ferrari D, Niks D, Yang LH et al (2003) Allosteric communication in the tryptophan synthase bienzyme complex: roles of the β-subunit aspartate 305−Arginine 141 Salt Bridge. Biochemistry 42(25):7807–7818

    Article  Google Scholar 

  6. Hao Y, Mingxin Y, Jiabin X, Lianqing Z, Tao Z, Zhihui Z (2019) Tongue squamous cell carcinoma discrimination with Raman spectroscopy and convolutional neural networks. Vib Spectrosc 103:102938

    Article  Google Scholar 

  7. Kim S, Park K, Sohn K et al (2016) Unified depth prediction and intrinsicdecomposition from a single image via joint convolutional neural fields. European Conference on Computer Vision. Springer International Publishing, pp 143–159

  8. Koller D, Friedman N (2009) Probabilistic graphical models: principles and techniques—adaptive computation and machine learning. MIT Press, London

    Google Scholar 

  9. Li Y, Brown MS (2014) Single image layer separation using relative smoothness. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp 2752–2759

  10. Liu L, Zhang J, Fu X et al (2019) Unsupervised segmentation and elm for fabric defect image classification. Multimedia Tools Appl 78(9):12421–12449

    Article  Google Scholar 

  11. Narihira T, Maire M, Yu SX (2015) Direct intrinsic Learning albedo-shadingdecomposition by convolutional regression. Proceedings of the IEEE International Conference on Computer Vision, pp 2992–2992

  12. Ren Z, Wu L (2018) Hyperspectral intrinsic image decomposition based on automatic subspace partitioning. Adv Laser Optoelectron 55(10):398–404

    Google Scholar 

  13. Roberto R-R, Edgar G, Ke P, Dang KN, Frédéric L, Philippe P, Lima-Saad WE (2019) Prediction of epileptic seizures with convolutional neural networks and functional near-infrared spectroscopy signals. Comput Biol Med 111:103355. https://doi.org/10.1016/j.compbiomed.2019.103355

    Article  Google Scholar 

  14. Sarishvili A, Winter J, Luhmann HJ, Mildenberger E (2019) Probabilistic graphical model identifies clusters of EEG patterns in recordings from neonates. Clin Neurophysiol 130(8):1342–1350

    Article  Google Scholar 

  15. Science—Computational Science; Reports on Computational Science Findings from University of Milan Provide New Insights (Efficient Computational Strategies To Learn the Structure of Probabilistic Graphical Models of Cumulative Phenomena)

  16. Scipioni M, Giorgetti A, Latta DD et al (2018) Direct parametric maps estimation from dynamic PET data: an iterated conditional modes approach. J Healthc Eng 4:1–14

    Article  Google Scholar 

  17. Shen Z, Yuan S (2019) Regional load clustering ensemble forecasting using convolutional neural network support vector regression machine. Power Grid Technol 10:15. https://doi.org/10.13335/j.1000-3673.pst.2019.0759

    Article  Google Scholar 

  18. Sun J, Yan H (2018) Research on expression classification method based on probability graph model. J Liaoning Univ Eng Technol (Natural Science Edition) 37(06):932–938

    Google Scholar 

  19. Sun L, Xie J, Wang C (2019) Collection development for macao studies—a user perspective. Collect Manage 44(2):1–15

    Google Scholar 

  20. Wang H (2016) Image-based visualization of plant leaf aging process. Shenyang Agricultural University

  21. Wang L, Zhong Y, Li Z, He Y (2019) On-line fabric defect detection algorithm based on in-depth learning. Comput Appl 1–6 [2019-04-01]

  22. Xu J (2016) Research on probabilistic diagnosis method of multi-fault program. Dalian Maritime University

  23. Xu J, Zhang D, Qian W (2017) Application of probabilistic graph model in social network user similarity discovery. Comput Sci Explor 11(07):1056–1067

    Google Scholar 

  24. Yang B (2015) Geometric feature extraction and shape restoration algorithm based on RGB-D image. Zhengzhou University

  25. Yang J (2016) Image fusion algorithm based on two-dimensional empirical mode decomposition. Northwest University of Technology

  26. Zhu R, Wei H, Lu Y, Sun D (2015) Research on license plate enhancement algorithm under uneven illumination. Minicomput Syst 36(03):601–604

    Google Scholar 

Download references

Acknowledgements

This work was supported by the National Key Research Development Program of China [grant number 2017YFB0802800].

Author information

Affiliations

Authors

Corresponding author

Correspondence to Zhisong Pan.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Li, X., Hu, G. & Pan, Z. An eigendecomposition method based on deep learning and probabilistic graph model. J Ambient Intell Human Comput 11, 3627–3637 (2020). https://doi.org/10.1007/s12652-019-01555-0

Download citation

Keywords

  • Eigendecomposition method
  • Deep learning
  • Probability graph model
  • Convolutional neural network
  • Conditional random field