350 rub
Journal Electromagnetic Waves and Electronic Systems №5 for 2014 г.
Article in number:
Face detection algorithm in case the additive white gaussian noise in image
Keywords: face detection support vector machines Sparse Network of Winnows additive white Gaussian noise
Authors:
L.A. Smaglit - Post-graduate Student, Yaroslavl State University. E-mail: shmaglit_lev@yahoo.com
A.L. Priorov - Dr. Sc. (Eng.), Associate Professor, Yaroslavl State University. E-mail: andcat@yandex.ru
V.V. Khryashchev - Ph. D. (Eng.), Associate Professor, Yaroslavl State University. E-mail: vhr@yandex.ru
D.V. Matveev - Master of Degree, Applied Mathematics and Computer Science, Yaroslavl State University. E-mail: yar_volley@inbox.ru
Abstract:
Face detection algorithms are applied in modern computer vision systems, robotics, video surveillance and access control interfaces of human-computer interaction. One of the most vital parameters is ability to face detect with high quality in case various distortions in image, which are strongly connected due to different noises in electronic devices. Therefore, the analysis of face detection algorithm quality in the presence of a complex signal-noses interaction is an actual issue. The additive white Gaussian noise in image was sought for the image distortion model. The comparison of three up-to-date face tracking approaches was done. Sparse Network of Winnows method showed the highest level of face detection - 87,3%. The second and the third results respectively have boosting (82,6%) method and support vector machines (62,4%). The additive white Gaussian noise impact on the quality of test methods was researched. The result of researches showcase that boosting algorithm is the most resistant to Gaussian noises. It should be to note that this approach outmarches the others on an average face detection level on over the (≈ 30%). An impact of pre-filtering procedure on the quality of face detection algorithms was investigated. Consequently, an average face detection level increase is achieved by the usage of Sparse Network of Winnows method on 31% whereas via the approach based on support vector machines on 10,5%. Proposed face detection algorithms can be applied for improvement the face detection probability in case various noises and distortions in digital image as well as for increase the robustness of identification, gender classification and non-related with faces detection methods.
Pages: 62-70
References

  1. Szeliski R. Computer Vision: Algorithms and Applications. Springer. 2010.
  2. Kriegman D., Yang M.H., Ahuja N. Detecting faces in images: A survey // IEEE Trans. on Pattern Analysis and Machine Intelligence. 2002. V.?24. № 1. P. 34−58.
  3. Hjelmas E. Face detection: A Survey // Computer vision and image understanding. 2001. V. 83. № 3. P. 236−274.
  4. Zhao W., Chellappa R., Phillips P., Rosenfeld A. Face recognition: A literature survey // ACM Computing Surveys (CSUR). 2003. V. 35. № 4. P. 399−458.
  5. Fasel B., Luettin J. Automatic facial expression analysis: A survey // Pattern Recognition Letters. 2003. V.?36. № 1. P. 259−275.
  6. Makinen E., Raisamo R. An experimental comparison of gender classification methods // Pattern Recognition Letters. 2008. V.?29. № 10. P. 1544−1556.
  7. Sammut C., Webb G.I. Encyclopedia of Machine Learning. Springer. 2011.
  8. Priorov A.L., Ignatov I.S., Golubev M.N., Xryashhev V.V. Razrabotka i analiz algoritmov vy'deleniya licz na izobrazheniyax // Proektirovanie i texnologiya e'lektronny'x sredstv. 2008. № 2. S. 58−62.
  9. Viola P., Jones M. Rapid object detection using a boosted cascade of simple features // Proc. Int. Conf. on Computer Vision and Pattern Recognition. 2001. № 1. P. 511−518.
  10. Nilsson M., Nordberg J., Claesson I. Face Detection Using Local SMQT Features and Split Up SNoW Classifier // Proc. of IEEE Int. Conf. ICASSP 2007. V.?2. P. 589−592.
  11. Kienzle W., Bakir G., Franz M., Schölkopf B. Face Detection - Efficient and Rank Deficient // Adv. in Neural Inf. Proc. Systems. 2005. V.?17. P. 673−680.
  12. Potapov A.S.Raspoznavanie obrazov i mashinnoe vospriyatie: obshhij podxod na osnove princzipa minimal'noj dliny' opisaniya. SPb.: Politexnika. 2007.
  13. Forsajt D.A., Pons D. Komp'yuternoe zrenie. Sovremenny'j podxod. M.: Vil'yams. 2004.
  14. Gonsales R., Vuds R. CZifrovaya obrabotka izobrazhenij. M.: Texnosfera. 2005.
  15. Tomasi C., Manduchi R. Bilateral Filtering for Gray and Color Images // Proc. Int. Conf. onComputerVisionandPatternRecognition. 1998. P. 839−846.
  16. Voloxov V., Sergeev E. Udalenie additivnogo belogo gaussovskogo shuma iz czifrovy'x izobrazhenij na osnove analiza glavny'x komponent // Tr. konf. GrafiKon-2010. 2010. S. 342−343.
  17. Freund Y., Schapire R. A decision-theoretic generalization of on-line learning and an application to boosting // Computational Learning Theory. 1995. P. 23−37.
  18. Miller D., Uyar H. A mixture of experts classifier with learning based on both labeled and unlabeled data // Neural information processing systems. 1996. № 9. P. 571−577.
  19. Nilsson M., Dahl M., Claesson I. The successive mean quantization transform // Proceedings of IEEE Int. Conf. ICASSP. 2005. V. 4. P. 429−432.
  20. Roth D., Yang M., Ahuja N. A SNoW-based face detector // in Advances in Neural Information Processing Systems 12 (NIPS 12). 2000. P. 855−861.
  21. Osuna E., Girosi F. Reducing the run-time complexity in support vector machines // Advances in Kernel Methods - SupportVector Learning. 1999. P. 271−284.
  22. Burges C. Simplified support vector decision rules // In International Conference on Machine Learning. 1996. P. 71−77.
  23. SHmaglit L.A., Golubev M.N., Ganin A.N., Xryashhev V.V. Gendernaya klassifikacziya po izobrazheniyu licza // 14‑ya mezhdunar. konf. «CZifrovaya obrabotka signalov i ee primenenie». M. 2012. T. 2. S. 425−428.
  24. Golubev M.N., SHmaglit L.A., Ganin A.N. Razrabotka i analiz sistemy' avtomaticheskogo raspoznavaniya pola lyudej po izobrazheniyu licza // XIV Vseros. nauchno-texn. konf. «Nejroinformatika‑2012». M. 2012. S. 151−159.
  25. Demidov M.A., Volchenkov N.G., Troiczkij A.K., Xramov A.A. Obnaruzhenie licz na izobrazheniyax na osnove poiska lokal'ny'x e'kstremumov // Informaczionno-izmeritel'ny'e i upravlyayushhie sistemy'. 2012. № 8. S. 55−61.