Article

Download article

Noise cleaning method for visual biometric information

Eugene Fedorov, Tetyana Utkina

Received 01.08.2021, Revised 05.11.2021, Accepted 20.12.2021

Abstract

One of the most important problems that exist in security systems today is to increase the effectiveness of identification of a person. Computer biometric identification speeds up and increases the accuracy of the recognition process, which is especially critical in a limited time. A special class of biometric identification of a person is formed by methods based on the analysis of visual information. The first step in processing the visual biometric information for analysis and subsequent recognition of objects such as a human face is digital image filtering or low-frequency noise elimination due to distortion of various imaging devices and their subsequent transmission through various communication channels. The paper proposes the noise cleaning method for visual biometric information by determining the structure of the smoothing filtering model of visual information about the identified person based on statistical evaluation of the noise cleaning quality of two-dimensional signal. The systematic analysis of modern noise cleaning methods for image is carried out. Smoothing adaptive linear time filtering; smoothing adaptive linear frequency filtering, called spectral subtraction; wavelet analysis with threshold processing; smoothing non-adaptive linear time filtering; smoothing non-linear filtration have been studied. It is established that the considered cleaning methods for visual biometric information have one or more of the following disadvantages: not automation of the choice of structure and parameters of the filter model and/or the low accuracy of additive and multiplicative noise cleaning. Therefore, it is important to develop the noise cleaning method of visual biometric information for pre-processing of human face images, which will ensure the necessary image quality and will not require time-consuming procedures to determine parameter values by the operator based on his empirical experience. The structure of the smoothing filtration model, which is reduced to determining the filter order, is determined. The characteristics and quality criterion of visual signal noise cleaning are offered. Numerous studies have been performed to determine the filter order parameter using the Siblings database, which allows to establish the most effective method based on statistical evaluation of the quality of visual information noise cleaning: in the case of additive Gaussian noise and in the case of multiplicative Gaussian noise, the least standard error, that meets the criterion of quality of visual signal noise cleaning, provides a medium -truncated filter. The proposed method allows to set and solve the problem of the visual signal pre-processing used for analysis and storage of visual information in intelligent computer systems of biometric identification of the person on the face image

Keywords:

visual signal; time filtering; frequency filtering; threshold processing; wavelet transform; additive Gaussian noise; multiplicative Gaussian noise

https://doi.org/10.24025/2306-4412.4.2021.247856

Retrieved from Volume 26, No. 4, 2021

Pages 5-15

References Suggested citation

References

[1] E. E. Fedorov, O. V. Nechyporenko, T. Yu. Utkina, and Ya. V. Korpan, Models and Methods of Computer Systems for Visual Image Recognition: a monograph. Cherkasy, Ukraine, 2021 [in Ukrainian].

[2] S. A. Broughton, and K. Bryan, Discrete Fourier Analysis and Wavelets. Applications to Signal and Image Processing. Hoboken, NJ, USA: John Wiley & Sons, 2018. 

[3] P. R. Hill, Audio and Speech Processing with MATLAB. Boca Raton, FL, USA: CRC Press, 2019. 

[4] A. K. Nandi, N. Sujatha, R. Menaka, and J. S. R. Alex, Computational Signal Processing and Analysis. Singapore: Springer, 2018. 

[5] W. Burger, and M. J. Burge, Digital Image. An Algorithmic Introduction Using Java. London, GB: Springer-Verlag, 2016. 

[6] K. S. Thyagarajan, Introduction to Digital Signal Processing using MATLAB with Application to Digital Communications. Cham, Switzerland: Springer, 2020. 

[7] B. Widrow, and S. Stearns, Adaptive Signal Processing. Moscow, Russia: Radio i svyaz, 1989 [in Russian].

[8] C. F. Cowan, and P. M. Grant, Adaptive Filters. Moscow, Russia: Mir, 1988 [in Russian].

[9] S. Haykin, Adaptive Filter Theory, Harlow, Essex, GB: Pearson Education, 2014. 

[10] P. S. R. Diniz, Adaptive Filtering Algorithms and Practical Implementation, Cham, Switzerland: Springer, 2020. 

[11] F. Gustafsson, Adaptive Filtering and Change Detection. Chichester, West Sussex, GB: John Wiley & Sons, 2000. 

[12] A. D. Poularikas, and Z. M. Ramadan, Adaptive Filtering Primer with MATLAB. Boca Raton, FL, USA: CRC Press, 2006. 

[13] M. Najim, Modeling, Estimation and Optimal Filtering in Signal Processing. Hoboken, NJ, USA: John Wiley & Sons, 2008. 

[14] M. G. Bellanger, Adaptive Digital Filters and Signal Analysis. New York, NY, USA: Marcel Dekker, 2001. 

[15] J. S. Lim, Two-Dimensional Signal and Image Processing. Englewood Cliffs, NJ, USA: Prentice Hall, 1990. 

[16] R. Gonzalez, R. Woods, and S. Eddins, Digital Image Processing in MATLAB. Moscow, Russia: Tekhnosfera, 2006 [in Russian].

[17] Methods of Computer Image Processing, V. A. Soyfer, Ed. Moscow? Russia: FIZMATLIT, 2003 [in Russian].

[18] N. Yu. Sekunov, PC Sound Processing. St. Petersburg, Russia: BHV-Petersburg, 2001 [in Russian].

[19] S. F. Boll, "Suppression of acoustic noise in speech using spectral subtraction", in Proc. IEEE Trans. on Acoustics, Speech, and Signal Processing (ASSP-27), 1979, vol .2, pp. 113-120.

[20] M. Berouti, R. Schwartz, and J. Makhoul, "Enhancement of speech corrupted by acoustic noise", in Proc. IEEE, Int. Conf. on Acoustics, Speech, and Signal Processing (ICASSP-79), 1979, pp. 208-211.

[21] R. J. McAulay, and M. L. Malpass, "Speech enhancement using a soft-decision noise suppression filter", in Proc. IEEE Trans. on Acoustics, Speech, and Signal Processing, 1980, vol. 28, no. 2, pp. 137-144.

[22] S. Malla, Wavelets in Signal Processing. Moscow, Russia: Mir, 2005 [in Russian].

[23] E. Fedorov, T. Utkina, O. Nechyporenko, and Y. Korpan, "Method of speech signal structuring and transforming for biometric personality identification", Communications in Computer and Information Science, vol. 1158, pp. 307-322, 2020.  doi: 10.1007/978-3-030-61656-4.

[24] G. X. Ritter, and J. N. Wilson, Handbook of Computer Vision Algorithms in Image Algebra. Boca Raton, FL, USA: CRC Press, 2001. 

[25] S.-T. Bow, Pattern Recognition and Image Preprocessing. New York, NY, USA: Marcel Dekker, 2002. 

[26] W. K. Pratt, Digital Image Processing. New York, NY, USA: John Wiley & Sons. 2001. 

[27] U. Qidwai, and C. H. Chen, Digital Image Processing: An Algorithmic Approach with MATLAB. Boca Raton, FL, USA: CRC Press, 2009. 

[28] N. N. Krasilnikov, Digital Processing of 2D and 3D Images. St. Petersburg, Russia: BHV-Petersburg, 2011 [in Russian]. 

[29] Mathematical Morphology and its Application to Image and Signal Processing, J. Goutsias, L. Vincent, and D. S. Bloomberg, Eds. Kluwer Academic Publishers, 2002. 

[30] Siblings Database. [Online]. Available: https://areeweb.polito.it/ricerca/cgvg/sibling sDB.html.

Suggested citation

Fedorov, E., & Utkina , T. (2021). Noise cleaning method for visual biometric information . Bulletin of Cherkasy State Technological University, 26(4), 5-15. https://doi.org/10.24025/2306-4412.4.2021.247856