An efficient approach for shadow detection based on Gaussian mixture model

Abstract

An efficient approach was proposed for discriminating shadows from moving objects. In the background subtraction stage, moving objects were extracted. Then, the initial classification for moving shadow pixels and foreground object pixels was performed by using color invariant features. In the shadow model learning stage, instead of a single Gaussian distribution, it was assumed that the density function computed on the values of chromaticity difference or bright difference, can be modeled as a mixture of Gaussian consisting of two density functions. Meanwhile, the Gaussian parameter estimation was performed by using EM algorithm. The estimates were used to obtain shadow mask according to two constraints. Finally, experiments were carried out. The visual experiment results confirm the effectiveness of proposed method. Quantitative results in terms of the shadow detection rate and the shadow discrimination rate (the maximum values are 85.79% and 97.56%, respectively) show that the proposed approach achieves a satisfying result with post-processing step.

Keywords

shadow detection / Gaussian mixture model / EM algorithm

References

[1] StaufferC, GrimsonW E L. Adaptive background mixture models for real-time tracking. [C]. IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Fort Collins, USA, 1999246-252

[2] StaufferC, GrimsonW E L. Learning patterns of activity using real-time tracking [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2000, 22(8): 747-757

[3] PratiA, MikicI, TrivediM M, CucchiaraR. Detecting moving shadows: algorithms and evaluation [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2003, 25(7): 918-923

[4] StauderJ, MechR, OstermannJ. Detection of moving cast shadows for object segmentation [J]. IEEE Transactions on Multimedia, 1999, 1(1): 65-76

[5] FungG S K, YungN H N, PangG K H, LaiA H S. Effective moving cast shadow detection for monocular color traffic image sequences [J]. Optical Engineering, 2002, 41(6): 1425-1440

[6] ChienS-y, HuangY-w, HsiehB-y, MaS-y, ChenL-gee. Fast video segmentation algorithm with shadow cancellation, global motion compensation, and adaptive threshold techniques [J]. IEEE Transactions on Multimedia, 2004, 6(5): 732-748

[7] XuD, LiX-l, LiuZ-k, YuanYuan. Cast shadow detection in video segmentation [J]. Pattern Recognition Letters, 2005, 26(1): 91-99

[8] XuD, LiuJ-z, LiX-l, LiuZ-k, TangX-ou. Insignificant shadow detection for video segmentation [J]. IEEE Transactions on Circuits and Systems for Video Technology, 2005, 15(8): 1058-1064

[9] HsiehJ-w, HuW-f, ChangC-j, ChenY-sheng. Shadow elimination for effective moving object detection by Gaussian shadow modeling [J]. Image and Vision Computing, 2003, 21(6): 505-516

[10] WangY, TanT, LoeK-f, JianK-wu. A probabilistic approach for foreground and shadow segmentation in monocular image sequences [J]. Pattern Recognition, 2005, 38(11): 1937-1946

[11] HsiehJ-w, YuS-h, ChenY-s, HuW-fong. Automatic traffic surveillance system for vehicle tracking and classification [J]. IEEE Transactions on Intelligent Transportation Systems, 2006, 7(2): 175-187

[12] LiZ, JiangP-h, MabH, YangJ, TangD-ming. A model for dynamic object segmentation with kernel density estimation based on gradient features [J]. Image and Vision Computing, 2009, 27(6): 817-823

[13] WangY, LoeK F, WuJ-kang. A dynamic conditional random field model for foreground and shadow segmentation [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2006, 28(2): 279-289

[14] JungC R. Efficient background subtraction and shadow removal for monochromatic video sequences [J]. IEEE Transactions on Multimedia, 2009, 11(3): 571-577

[15] LiZ-y, LiuC-c, LiuG-h, YongC, YangX-b, ZhaoC-rong. A novel statistical image thresholding method [J]. Aeu-International Journal of Electronics and Communications, 2010, 64(12): 1137-1147

[16] WangYang. Real-time moving vehicle detection with cast shadow removal in video based on conditional random field [J]. IEEE Transactions on Circuits and Systems for Video Technology, 2009, 19(3): 437-441

[17] FangL-z, QiongW-y, ShengY-zhi. A method to segment moving vehicle cast shadow based on wavelet transform [J]. Pattern Recognition Letters, 2008, 29(16): 2182-2188

[18] LinC-t, YangC-t, ShouY-w, ShenT-kuei. An efficient and robust moving shadow removal algorithm and its applications in ITS [J]. Eurasip Journal on Advances in Signal Processing, 2010, 2010: 1-19

[19] CucchiaraR, PiccardiM, PratiA, SirottiS. Improving shadow suppression in moving object detection with HSV color information [C]. 2001 IEEE Intelligent Transportation Systems Proceedings. Oakland (CA), USA, 2001334-339

[20] CucchiaraR, GranaC, PiccardiM, PratiA. Detecting moving objects, ghosts, and shadows in video streams [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2003, 25(10): 1337-1342

[21] SalvadorE, CavallaroA, EbrahimiT. Cast shadow segmentation using invariant color features [J]. Computer Vision and Image Understanding, 2004, 95(2): 238-259

[22] BenedekC, SziranyiT. Study on color space selection for detecting cast shadows in video surveillance [J]. International Journal of Imaging Systems and Technology, 2007, 17(3): 190-201

[23] NadimiS, BhanuB. Physical models for moving shadow and object detection in video [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2004, 26(8): 1079-1087

[24] NicolasM B, ZaccarinA. Learning and removing cast shadows through a multidistribution approach [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2007, 29(7): 1133-1146

[25] ZhangW, FangX-z, YangX K K, WuQ M J. Moving cast shadows detection using ratio edge [J]. IEEE Transactions on Multimedia, 2007, 9(6): 1202-1214

[26] AmatoA, MozerovM G, BagdanovA D, GonzalezJ. Accurate moving cast shadow suppression based on local color constancy detection [J]. IEEE Transactions on Image Processing, 2011, 20(10): 2954-2966

[27] ChoiJ-m, YooY-j, ChoiJ-young. Adaptive shadow estimator for removing shadow of moving object [J]. Computer Vision and Image Understanding, 2010, 114(9): 1017-1029

[28] RednerR A, WalkerH F. Mixture densities, maximum-likelihood and the EM algorithm [J]. SIAM Review, 1984, 26(2): 195-237

[29] MikicI, CosmanP C, KogutG T, TrivediM M. Moving shadow and object detection in traffic scenes [C]. 15th International Conference on Pattern Recognition, 2000321-324

[30] HorprasertT, HardwoodD, DavisL S. A statistical approach for real-time robust background subtraction and shadow detection [C]. The Proceedings of the Seventh IEEE International Conference on Computer Vision. Corfu, Greece, 1999898-903