Fast Shadow Detection and Removal Based on Saliency

Abstract:

In this paper, we present a new method for removing shadows from images. Different from traditional methods that explore pixel or edge information, we employ a shadow saliency region based approach. We combine the region saliency with the clustering algorithm to get shadow region with the entropy rate in an unsupervised manner. The shadow free image is recovered by relighting each pixel based on our lighting model, and any visible shadow boundaries in the image can be eliminated by the Bayesian framework. We show that our method is fast and can generate satisfactory results for images with shadows.

• Periodical: Applied Mechanics and Materials (Volume 703)
• Pages: 286-289
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.703.286
• Online since: December 2014
• Authors: Qi Chen*, Xing Ben Yang, Shao Wen Yang
• Keywords: Entropy Rate, Saliency Estimation, Shadow Removal

References

[1] R. Guo, Q. Dai, D. Hoiem. In CVPR (2011), IEEE. 2033–(2040).

[2] C. Xiao, N. Yongwei. Visualization and Computer Graphics, IEEE Transactions on (2011)99, 1135–1147.

[3] G.D. Finlayson,S. D, Hordley C. Lu and M.S. Drew. IEEE Trans on PAMI (2006). 28. 59-68.

[4] J. Stauder, R. Mech, J. Osterman. IEEE Transactions on Multimedia (1999), 65–76.

[5] N.C. Foley,S. Grossberg,E. Mingolla. Cognitive Psychol(2012)6, 577–117.

[6] F. Perazzi, P. Krahenbuhl, Y. Pritch, A. Hornung. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (2012 ), 733–740.

DOI: 10.1109/cvpr.2012.6247743

[7] P. Jiang, H. Ling, J. Yu, and J. Peng. In ICCV(2013).

[8] C.X. Xiao, Y. Sher, D. L. XIAO, L. MAK. In Computer Graphics Forum (2013).

[9] J. Zhu, G. Samuelk, S. Z. Masood, F. Tappenm. In CVPR (2010).

[10] Available from http: /www. cs. sfu. edu/~colour/data.

[11] S. Bianco and R. Schettini. In CVPR (2012), 65–72.

[12] A. Gijsenij, R. Lu, and T. Gevers. Image Processing, IEEE Transactions on (2012): 697–707.