Active-GNG: model acquisition and tracking in cluttered backgrounds

ABSTRACT

The Self-Organising Artificial Neural Network Models, of which we have used the Growing Neural Gas (GNG) can be applied to preserve the topology of an input space. Traditionally these models neither do include local adaptation of the nodes nor colour information. In this paper, we extend GNG by adding an active step to the network, which we call Active-Growing Neural Gas (A-GNG) that has both global and local properties and can track in cluttered backgrounds. The approach is novel in that the topological relations of the model are based on a number of attributes (e.g. global and local transformations, mapping function and skin colour information) which allow us to automatically model and track 2D gestures. To measure the quality of the tracked correspondences we use two interlinked topology preservation measures. Experimental results have shown better performance of our proposed method over the original GNG and the Active Contour Model.

REFERENCES

• 1: A. Angelopoulou, A. Psarrou, G. Gupta, and J. García. Nonparametric Modelling and Tracking with active-GNG. IEEE Workshop on Human Computer Interaction, HCI 2007, in conjuction with ICCV 2007, LNCS 4796, pages 98--107, 2007.
• 2: H. Bauer and K. Pawelzik. Quantifying the neighbourhood preservation of self-organizing feature maps. IEEE Transactions on Neural Networks, 3(4):570--579, 1992.
• 3: Adam Baumberg , David Hogg, Learning flexible models from image sequences, Proceedings of the third European conference on Computer vision (vol. 1), p.299-308, June 1994, Stocklholm, Sweden
• 4: Christopher M. Bishop, Neural Networks for Pattern Recognition, Oxford University Press, Inc., New York, NY, 1995
• 5: T. F. Cootes , C. J. Taylor , D. H. Cooper , J. Graham, Active shape models—their training and application, Computer Vision and Image Understanding, v.61 n.1, p.38-59, Jan. 1995 [doi>10.1006/cviu.1995.1004]
• 6: D. Cremers, T. Kohlberger, and C. Schnorr. Shape statistics in kernel space for variational image segmentation. Pattern Recognition, 36(9):1929--1943, 2003.
• 7: H. R. Davies, J. C. Twining, F. T. Cootes, C. J. Waterton, and J. C. Taylor. A Minimum Description Length Approach to Statistical Shape Modeling. IEEE Transaction on Medical Imaging, 21(5):525--537, 2002.
• 8: Nicolae Duta , Anil K. Jain , Marie-Pierre Dubuisson-Jolly, Automatic Construction of 2D Shape Models, IEEE Transactions on Pattern Analysis and Machine Intelligence, v.23 n.5, p.433-446, May 2001 [doi>10.1109/34.922703]
• 9: F. Florez-Revuelta, J. M. Garcia-Chamizo, J. Garcia-Rodriguez, and A. Hernandez-Saez. Geodesic topographic product: An improvement to measure topology preservation of self--organizing neural networks. Advances in Artificial Intelligence-IBERAMIA 2004, 3315:841--850, 2004.
• 10: B. Fritzke. A growing Neural Gas Network Learns Topologies. In Advances in Neural Information Processing Systems 7 (NIPS'94), pages 625--632, 1995.
• 11: Shen Furao , Osamu Hasegawa, An incremental network for on-line unsupervised classification and topology learning, Neural Networks, v.19 n.1, p.90-106, January 2006 [doi>10.1016/j.neunet.2005.04.006]
• 12: Andrew Hill , Chris J. Taylor , Alan D. Brett, A Framework for Automatic Landmark Identification Using a New Method of Nonrigid Correspondence, IEEE Transactions on Pattern Analysis and Machine Intelligence, v.22 n.3, p.241-251, March 2000 [doi>10.1109/34.841756]
• 13: M. Kass, A. Witkin, and D. Terzopoulos. Snakes: Active Contour Models. In Proc. of the 1st Internationl Conference on Computer Vision, IEEE Computer Society Press, pages 259--268, 1987.
• 14: S. Lankton, D. Nain, A. Yezzi, and A. Tannenbaum. Hybrid geodesic region-based curve evolutions for image segmentation. In Proc. SPIE: Med. Img., 6510:65104U, 2007.
• 15: John MacCormick , Michael Isard, Partitioned Sampling, Articulated Objects, and Interface-Quality Hand Tracking, Proceedings of the 6th European Conference on Computer Vision-Part II, p.3-19, June 26-July 01, 2000
• 16: C. Nastar and N. Ayache. Fast segmentation, tracking and analysis of deformable objects. In Proc. of the 4th International Conference on Computer Vision, ICCV'93, pages 275--279, 1993.
• 17: L. H. Staib and J. S. Duncan. Parametrically deformable contour models. In IEEE conference on Computer Vision and Pattern Recognition, pages 427--430, 1989.
• 18: E. B. Sudderth, M. I. Mandel, W. T. Freeman, and A. S. Willsky. Distributed Occlusion Reasoning for Tracking with Nonparametric Belief Propagation. In Advances in Neural Information Processing Systems, 17:1369--1376, 2005.
• 19: Alan L. Yuille , Peter W. Hallinan , David S. Cohen, Feature extraction from faces using deformable templates, International Journal of Computer Vision, v.8 n.2, p.99-111, Aug. 1992 [doi>10.1007/BF00127169]

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.4 IMAGE PROCESSING AND COMPUTER VISION
      I.4.8 Scene Analysis
          Subjects: Shape

Additional Classification:
  I. Computing Methodologies
  I.4 IMAGE PROCESSING AND COMPUTER VISION
      I.4.8 Scene Analysis
          Subjects: Tracking
  I.5 PATTERN RECOGNITION
      I.5.1 Models
          Subjects: Neural nets

General Terms:
Algorithms, Experimentation, Theory

Keywords:
nonrigid shapes, self-organising networks, tracking, unsupervised learning