Comprehensive method for restoring graphic data based on morphological algorithms.
Abstract
Modern methods of graphic data recovery based on mathematical modeling of noise distribution and optical aberrations, determination of the threshold value, the use of morphological methods, as well as neural network algorithms are considered. The priority of such approaches as development of mathematical models of statistical distribution, optimization of the threshold value, development of morphological methods for recovery of the color image matrix, application of wavelet transformation, as well as adjustment of the neural network architecture of the autoencoder and stacked autoencoder is indicated. A technique has been developed based on the application of software algorithms for processing an image matrix, which includes the selection of homogeneous areas using the method of connected principal and the subsequent restoration of graphic data through the implementation of erosion and dilatation procedures. The development of a system for evaluating the image matrix restoration effectiveness includes the concepts of objective functions and arguments of objective functions, as well as system constraints that can be formulated in accordance with the given task, which made it possible to reduce the task of setting up and optimization to the task of finding extrema of a function. The objective functions of restoring the image matrix by software algorithms include the average error of restoring the array of graphic data and the corresponding value for the auxiliary channel, as well as the total time of performing a set of operations for restoring the array of graphic data. At the same time, the arguments of the objective functions consist of the parameters of the correction primitive, the minimum permissible deviation for each color channel within a separate segment, and the number of iterations calculated accordingly for the given task. It is indicated that this approach can be adapted to the color scheme, according to which graphical data are presented (dynamic range, presence of an auxiliary channel). At the same time, the developed technique is characterized by a simple mathematical apparatus and a minimum load on the computing resource.
References
2. Ye, H. Y. H., Gao, Z. G. Z., Qin, Z. Q. Z., & Wang, Q. W. Q. (2013). Near-infrared fundus camera based on polarization switch in stray light elimination. Chinese Optics Letters, 11(3), 031702–31705.
3. Prabhakara, R. S., Wright, C. H. G., & Barrett, S. F. (2012). Motion Detection: A Biomimetic Vision Sensor Versus a CCD Camera Sensor. IEEE Sensors Journal, 12(2), 298–307.
4. Pankaj, D., Narayanankutty, K., & Govind, D. (2018). Image Denoising Using Total Variation Wavelet Galerkin Method. Procedia Computer Science, 143, 481-492.
5. Jing, X. J., & Min, Y. (2015). An linearized alternating direction method for total variation image restoration. The 27th Chinese Control and Decision Conference (2015 CCDC).
6. Lamei, Z., & Sheng, C. (2014). A method for PET image denoising based on total variation minimization. Advanced Computer Control.
7. Sebe, N. (2001). Improving visual matching: Similarity noise distribution and optimal metrics. S.l.: Thesis PhD Computer Science, Leiden University.
8. Mahajan, V. N. (2014). Aberration theory made simple. Volume: TT93. SPIE Press.
9. Guruprasad, P. (2020). Overview of different thresholding methods in image processing. Conference: TEQIP Sponsored 3rd National Conference on ETACC. Project: Word image identification for handwritten scripts using data visualization.
10. Pankaj, D., Narayanankutty, K., & Govind, D. (2018). Image Denoising Using Total Variation Wavelet Galerkin Method. Procedia Computer Science, 143, 481-492.
11. Shih, F.Y. (2017). Image processing and mathematical morphology: Fundamentals and applications. CRC Press.
12. Goutsias, J., Vincent, L. M., & Bloomberg, D. S. (2002). Mathematical morphology and its applications to image and signal processing. Boston: Kluwer Academic.
13. M., R. J., & Andreadēs, I. (2000). Special issue Mathematical morphology & nonlinear image processing. Oxford u.a.: Pergamon.
14. Suzuki, Y., & Ozaki, T. (2017). Stacked Denoising Autoencoder-Based Deep Collaborative Filtering Using the Change of Similarity. 2017 31st International Conference on Advanced Information Networking and Applications Workshops (WAINA).
15. Ye, X., Wang, L., Xing, H., & Huang, L. (2015). Denoising hybrid noises in image with stacked autoencoder. 2015 IEEE International Conference on Information and Automation.
16. Starck, J.-L., Murtagh, F., & Fadili, J. M. (2010). Sparse image and signal processing: Wavelets, curvelets, morphological diversity. Cambridge University Press.
17. Guo, H., Ono, N., & Sagayama, S. (2008). A structure-synthesis image inpainting algorithm based on morphological erosion operation. 2008 Congress on Image and Signal Processing.
18. Dougherty, E. R., & Lotufo, R. A. (2003). Hands-on morphological image processing. SPIE Optical Engineering Press.
19. Hÿtch Martin, & Hawkes, P. W. (2020). Morphological image operators. Academic Press, an imprint of Elsevier.
Abstract views: 83 PDF Downloads: 76
