Video Object Classification System with Shadow Removal using Gaussian Mixture Model
Keywords:Classification, Gaussian Mixture Model, background subtraction, Maximum A Posteriori and Shadow detection
Classification is the process of assigning a class to a group of objects. Moving objects classification can be difficult a task in the presence of dynamic factors like occlusion clutters and shadows. This paper developed a classifier for moving images (video stream) by using a modified adaptive background mixture model method. This system removes shadows and correctly classifies moving objects as human, human group and vehicles.
Background Mixture Model is common technique used in Computer Vision, Video object classification is not an exception, many background models have been designed to address different problems ranging from slow start and shadow removal; this paper presents a method which models each pixel as a mixture of Gaussians and using a Maximum A Posteriori approximation to update the model, this paper also introduces a two level shadow removal technique which suppressed shadows in colour and texture consistencies in the classified objects so that the system will not mis- classify moving shadows as objects. This work overcome the problem of slow learning in busy environment and can classify more than one object in view of the camera
(1) Cutler, R. and Davis, L.S. (2000): “Robuct real-time motion detection, analysis and Applications”. In IEEE Transactions on Pattern Analysis and Machine Intelligence, volume 8, pp. 781-796.
(2) Dahmen, J., Schluter, R., and Ney, H. (1999): “Discriminative training of Gaussian Mixtures for image Object Recognition”. In proceeding 21st Symposium of german Association for pattern
Recognition, Bonn, germany, pp. 205-212.
(3) Dedeoglu Y.(2004): Moving object detection, tracking and classification for smart video surveillance,Master's Thesis, Bilkent University, Ankara,pp2.
(4) Duda, R.O. and Hart, P.E. (2002): “Pattern Classification and Scene Analysis”. Nevada. John Wiley and Sons, inc. pp. 123-124.
(5) Haritaoglu I.,harwood L.and Davis S. (1998): “W4 Who? When?where? what? A real time System for detecting and tracking people” FGR98 submitted.
(6) KaewTrakulPong, P. and Bowden, R. (2001): “An improved Adaptive Background Mixture Model for Real-Time Tracking with Shadow Detection”. In proceeding. 2nd European Workshop on Advanced Video based Surveillance Systems.
(7) Landabaso J-L,Pardàs M and Xu L.(2011): Shadow Removal with Morphological Reconstruction Proceedings of the 17th International Conference on Pattern Recognition, Vol. 3, pp.810-813
(8) Omidiora,E.O., Fakolujo,O.A., Ayeni,R.O., Olabiyisi,S.O. and Arulogun O.T. (2008) “Quantittive Evaluation of PCA and FDA in Teaching of face Image” Journal of Computer Science and its Application 15(1), pp22-37
(9) Phong, B.T. (1975): Illumination for computer generated pictures, Communications of ACM Volume (18) 6, Pp. 311–317.
(10) Prati A.,Mikic I.,Trivedi M. and Cucchiara R.(2003):Detecting
Moving Shadows: algorithm and Evaluation.IEEE Transactions on Pattern Analysis and Machine Intelligence, 25(7):Pp.918-923
(11) Reynolds D,Ridder C., Munkelt O. and Kirchner H.(1995): Adaptive background estimation and foreground detection using kalman-filtering. Proceedings of the International Conference on Recent Advances in Mechatronics,Academic Press, pp: 193-199.
(12) Stauffer C. and Grimson W. (1998):” Adaptive background mixture models for real-time Tracking”. Proc. Of the IEEE Computer Society conference on Computer Vision and pattern recognition, pp 246-252.
(13) Stauffer C. and Grimson W. E. L.(1999) “Adaptive background mixture models for real-time tracking”. Proc. Of the IEEE Computer Society conference on Computer Vision and pattern recognition (Cat. No
PROO149). IEEE Comput. Soc. Part Vol. 2.pp 87-94
(14) Stauffer, C. and Grimson, W. E. L. (2000). Learning patterns of activity using real-time tracking.IEEE Transaction on Pattern Analysis and Machine Intelligence, Vol.22(8):pp.747–757.
(15) Wixson L. and Selinger A. (1998):” Classifying moving objects as rigid or non-rigid”. In Proceeding of DARPA Image Understanding Workshop, pp. 341-358.