Video Dizilerinde Hareket Tanıma

Bu çalışmada, video dizilerinden insan hareketlerinin tanınması için yapay sinir ağı (YSA) ve saklõmarkov modelini (SMM) temel alan insan hareket tanıma sistemleri modellenmiştir. Hareket tespit ve tanıma sistemi üç aşama içermektedir. Birinci aşamada, video çerçevelerinde insan pozu tespit edilmektedir. ikinci aşamada, hareketlere ait poz dizileri elde edilmektedir. Üçüncü aşamada, YSA ve SMM tabanlı tanıma modelleri hareket tanıma için kullanılır. Önerilen YSA modeli yüksek tanıma oranlarına sahiptir. Bununla birlikte, SMM modeli için yeni eylemler önceki eğitimli SMM’lerde herhangi bir değişiklik yapmadan kolayca eklenebilir.

Anahtar Kelimeler:

hareket tanıma, poz sembolü, pozdizisi, ysa, smm

ACTION RECOGNITION IN VIDEO SEQUENCE

In this study, human activity recognition systems based on Artificial Neural Networks (ANN) and Hidden Markov Model (HMM) are modeled for recognition of the human activities from video sequences. The action recognition system models consist of three stages. At the first stage, human pose is detected in video frames. At the second stage, the pose sequences are obtained. At the third stage, ANN, and HMM based recognition models are used for action recognition. Sugested ANN model has higher recognition rate. However, for the HMM models new actions can be easily added without making any changes in the previous trained HMM.

Keywords:

Action Recognition, Pose Symbols, Pose Sequence, ANN, HMM..,

PDF

___

[1] M. H. Yangand, N. Ahuja, “Recognizing hand gesture using motion trajectories”, Proceedings of the IEEE Conference on Computer Vision and Image Understanding, Fort Collins, Colorado, June 1999, pp. 468-472.
[2] C. Myers, L. Rabinierand A. Rosenberg, ‘Performance tradeoffs in dynamic time warping algorithms for isolated word recognition’, IEEE Trans. on Audio Speech Lang. Process., pp.623635, 1980.
[3] C. Rao, A. Yilmazand M. Shah, “View-invariant representation and recognition of actions”, Int. J.. of Computer Vision, Vol. 50, No.2, pp. 203-226, 2002.
[4] N. Cuntoor, A. Kale and R. Chellappa, ‘Combining multiple evidences for gait recognition’, Proceedings of International Conference on Acoustics, Speech and Signal Processing, 2003, Vol. 3, pp. 113-116.
[5] D. Cunado, M. S. Nixonand J. N. Carter, ‘Using gait as a biometric, via phase-weighted magnitude spectra,’ Proceedings of 1st International Conference on Audio and Video Based Biometric Person Authentication, 1997, pp. 95-102.
[6] H. Fujiyoshi, A. J. Liptonand T. Kanade, “Real time human motion analysis by õmage skeletonization”, IEICE Trans Inf Sys, pp. 113-120, Vol. E87-D, No.1, 2004.
[7] A. F. Bobickand J. W. Davis, “The recognition of human movementusing temporal templates”, IEEE Trans. on Pattern Anal. Mach. Intell., Vol. 23 (3), pp. 257-267, 2001.
[8] R. T. Collins, R. Grossand J. Shi, ‘Silhouette based human identification from body shapeand gait’, 5th International Conference on Automatic Face and Gesture Recognition, May 2002.
[9] J. Yamato, J. Ohyaand K. Ishii, “Recognizing human action in time sequential images using hidden markov models”, Proceedings Computer Vision and Pattern Recognition, 1992, pp. 379385.
[10] C. Bregler, “Learning and Recognizing Human Dynamics in Video Sequences’, Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, June 1997, pp. 568-574.
[11] A. Kale, A.N. Rajagopalan, N. Cuntoorand V. Krueger, ‘Gait based Recognition of Humans Using Continuous HMMs’, Proceedings of the IEEE International Conference on Automatic Face and Gesture Recognition, May 2002, pp. 336-341.
12] Y. A Ivanovand A. F. Bobick, “Recognition of visual activities and õnteractions by stochastic parsing”, IEEE Trans. on Pattern Anal. Mach. Intell.,Vol.22, No.8, pp.852-872, 2000.
[13] X. Feng and P. Perona,“Human action recognition by sequence of movelet code words”, Proceedings of 1th International Symposium on 3D Data Processing Visualization and Transmission, 2002, pp. 717–721.
[14] H. S. Chen, H. T. Chen, Y. W. Chen, and S. Y. Lee, “Human action recognition using star skeleton”, Proc. of 4th ACM International Workshop on Video Surveillanceand Sensor Networks 2006 (VSSN-2006) ,in conjunction with ACM Multimedia 2006, Santa Barbara, CA, USA, October 27, 2006.
[15] K. Hatun, P. Duygulu, “A new representation for action recognition using sequence of posewords”, 19th International Conference on Pattern Recognition, 2008.
[16] Güçlü T., Adar N., “Recognizing human actions by using visual posewords”, Int. Symposium on Innovations in Inteligent Systems and Applications (INISTA 2010), Haziran 2010 Kayseri.
[17] Kandemir C.M., “Yüksek Baúarõmlõ, Bilgisayarla Görü Uygulamalarõ Programlamasõ”, Doktora Tezi, Eskiúehir Osmangazi Üniversitesi Fen Bilimleri Enstitüsü, 125s., 2009.
[18] D. Ramanan, C. Sminchisescu, “Training deformable models for localization”, IEEE Conference on Computer Vision and Pattern Recognition, 2006, Vol. 1, 206–213.
[19] T. Koga, K. Iinuma, A. Hirano, Y. Iijima, and T.Ishiguro, “Motion compensated interframe coding for video conferencing,” Proc. Nat. Telecommun. Conf., New Orleans, LA, Nov. 29– Dec. 3 1981, pp. G5.3.1–G5.3.5.
[20] H. Demuth, M. Beale, Neural Network Toolbox for use with MATLAB, 2002.