350 rub
Journal Neurocomputers №2 for 2010 г.
Article in number:
Facial image segmentation based on mixed color space
Keywords: facial image segmentation color spaces color appearance models
Authors:
S. I. Anishenko, D. G. Shaposhnikov, R. Comley, X. Gao
Abstract:
Currently, one of the major problems in the development of segmentation algorithms for systems of human-computer interaction lies in the fact that such systems should work under different lighting conditions, a wide range of head rotations, as well as for people with different skin color. In this work, an analysis of the color spaces and color appearance models is performed in order to select their components, which are more suitable for skin color modeling and invariant to illuminance and race. On the all analyzed images (n=90) a facial areas were marked up by hand. In that areas the histograms of all attributes of color spaces (HSV, HSL, RGB, XYZ) and color appearance models (CIELAB, CIELUV, CIECAM02) were calculated and studied. On each histogram the maximum peak was found and considered as the most probable value of the attribute in facial areas. It was shown that most stable attributes, under changed external conditions of capturing, were the hue (H) of space HLS, saturation (S), achromatic response (A), colourfulness (M) and chroma (C) of the model CIECAM02. The values of those components for skin color satisfy the conditions: 3 < H < 13, 18 < A < 28, 15 < M-S < 18, 10 < C-S < 13. Image pixels which fall into the specified range, united in the potential areas of interest and verified by the edge density of the subjective brightness (Q, model CIECAM02). For the face area the value of the density lies in the range 0.15 <<0.5. Developed algorithm has been tested using image database (n=180) captured under room lighting condition (illuminance 40-80 lux). Using the selected color attributes and the subsequent verification of the edges density in the segmented areas, a detection of faces was correct in 94.5% of cases. Besides, the developed algorithm, unlike many known methods, has demonstrated invariance with respect to human race, illuminance and head pose (rotation up to 80°).
Pages: 40-46
References
  1. Jaimes, A. and Sebe, N., Multimodal human-computer interaction: A survey // Computer Vision and Image Understanding. Elsiver. 2007. V.108. iss.1-2. P.116-134.
  2. Poole, A. Ball, L. J., Eye tracking in human-computer interaction and usability research: current status and future prospects // In Ghaoui, Claude (Ed.). Encyclopedia of Human Computer Interaction. Idea Group, http://www.alexpoole.info (in press).
  3. Murphy-Chutorian, E., Trivedi, M., Head pose estimation in computer vision: A Survey // IEEE Transactions on Pattern Analysis and Machine Intelligence 2009. V. 31. No. 4. P. 607-626.
  4. Gizatdinova, Y., Surakka, V., Feature-based detection of facial landmarks from neutral and expressive facial images // IEEE Transactions on Pattern Analysis and Machine Intelligence. 2006. V. 28. No. 1. P. 135-139.
  5. Hjelmas, E., Low, B., Face detection: A survey // Computer Vision and Image Understanding. 2001. V. 83. P. 235-274.
  6. Huang, Ch., Ai, H., Li, Y., and Lao, Sh., High-performance rotation invariant multiview face detection // IEEE Transactions on Pattern Analysis and Machine Intelligence. 2007. V. 29. No. 4. P. 671-686.
  7. Picard, R., Affective Computing. M.I.T. Press 1997. Cambridge, MA.
  8. Stan, Z., Li, Anil Jain, K., Handbook of face recognition. Springer. New York. USA. 2005.
  9. Viola, P. and Jones, M., Robust real-time face detection // International Journal of Computer Vision. 2004. V. 57. No. 2. P. 137-154.
  10. Wang, P., Ji, Q., Multi-view face and eye detection using discriminant features // Computer Vision and Image Understanding. 2007. V. 105. No. 2. P. 99-111.
  11. Cai, J., and Goshtasby, A., Detecting human faces in color images // Image and Vision Computing 1999. V. 18. P. 63-75.
  12. Shin, M. C., Chang, K. I., and Tsap, L. V., Does colorspace transformation make any difference on skin detection - // In Proc. 6th IEEE Workshop on Applications of Computer Vision. 2002. P. 275-279.
  13. Terrillon, J.-C., Shirazi, M. N., Fukamachi, H., and Akamatsu, S., Comparative performance of different skin chrominance models and chrominance spaces for the automatic detection of human faces in color images // In Proc. International Conference on Face and Gesture Recognition. 2000. P. 54-61.
  14. Vezhnevets, V., Sazonov, V., and Andreeva, A., A Survey on pixel-based skin color detection techniques // Proc. Graphicon-2003. Moscow. Russia. Sept. 2003. P. 85-92.
  15. Фершильд М. Д. Модели цветового восприятия, 2-я редакция. Рочестер: Рочерстерскийтехнологическийинститут. 2004.
  16. CIE, The CIE 1997 Interim Colour Appearance Model (Simple Version), CIECAM97s, CIE TC1-34, April, 1998.
  17. Gao, X., Hong, K., Podladchikova, L., Shaposhnikov, D.and Passmore, P., Colour appearance based approaches for segmentation of traffic signs // EURASIP Journal on Image and Video Processing. ArticleID 386705, 2008. doi:10.1155/2008/386705.
  18. Шапошников Д. Г., Головань А. В., Подладчикова Л. Н., Шевцова Н.А., Гао К., Гусакова В. И., Гизатдинова Ю. Ф. Применение поведенческой модели зрения для инвариантного распознавания лиц и дорожных знаков // Нейрокомпьютеры: разработка и применение. 2002. № 7-8. С. 21-33.
  19. Павлидис Т. Алгоритмы машинной графики и обработки изображений: Пер. с англ. - М.: Радио и связь. 1986.
  20. Anishenko, S., Osinov, V., Shaposhnikov, D., Podladchikova, L., Comley, R., Sukholentsev, K., and Gao, X., A Motion Correction System for Brain Tomography Based on Biologically Motivated Models // Proc. 7-th IEEE Int. Conf. On Cybernetic Intelligent Systems, UK. 2008. P. 32-36.