Memetic Computing

, 1:291 | Cite as

A vision architecture for unconstrained and incremental learning of multiple categories

  • Stephan Kirstein
  • Alexander Denecke
  • Stephan Hasler
  • Heiko Wersing
  • Horst-Michael Gross
  • Edgar Körner
Special Issue - Regular Research Paper


We present an integrated vision architecture capable of incrementally learning several visual categories based on natural hand-held objects. Additionally we focus on interactive learning, which requires real-time image processing methods and a fast learning algorithm. The overall system is composed of a figure-ground segregation part, several feature extraction methods and a life-long learning approach combining incremental learning with category-specific feature selection. In contrast to most visual categorization approaches, where typically each view is assigned to a single category, we allow labeling with an arbitrary number of shape and color categories. We also impose no restrictions on the viewing angle of presented objects, relaxing the common constraint on canonical views.


Real-time vision system Life-long learning Category learning Vector quantization Hand-held objects 


  1. 1.
    Agarwal S, Awan A, Roth D (2004) Learning to detect objects in images via a sparse, part-based representation. IEEE Trans Pattern Anal Mach Intell 26(11): 1475–1490CrossRefGoogle Scholar
  2. 2.
    Arsenio AM (2004) Developmental learning on a humanoid robot. In: Proceedings of the international joint conference on neuronal networks (IJCNN), pp 3167–3172Google Scholar
  3. 3.
    Denecke A, Wersing H, Steil JJ, Körner E (2009) Online figure-ground segmentation with adaptive metrics in generalized LVQ. Neurocomputing 72(7–9): 1470–1482CrossRefGoogle Scholar
  4. 4.
    French RM (1999) Catastrophic forgetting in connectionist networks. Trends Cognit Sci 3(4): 128–135CrossRefMathSciNetGoogle Scholar
  5. 5.
    Fritsch J, Lang S, Kleinehagenbrock M, Fink GA, Sagerer G (2002) Improving adaptive skin color segmentation by incorporating results from face detection In: Proceedings of the IEEE International workshop on robot and human interactive communication (ROMAN), Berlin, pp 337–343Google Scholar
  6. 6.
    Fritzke B (1994) Growing cell structures—a self-organizing network for unsupervised and supervised learning. Neural Networks 7(9): 1441–1460CrossRefGoogle Scholar
  7. 7.
    Fritzke B (1995) A growing neural gas network learns topologies. In: Tesauro G, Touretzky DS, Leen TK (eds) Advances in Neural Information Processing Systems 7. MIT Press, Cambridge, pp 625–632Google Scholar
  8. 8.
    Fritz M, Kruijff G-JM, Schiele B (2007) Cross-modal learning of visual categories using different levels of supervision. In: Proceedings of the international conference on vision systems (ICVS)Google Scholar
  9. 9.
    Fukushima K (1980) Neocognitron: a self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position. Biol Cybern 36(4): 193–202zbMATHCrossRefGoogle Scholar
  10. 10.
    Furao S, Hasegawa O (2006) An incremental network for on-line unsupervised classification and topology learning. Neural Netw 1(19): 90–106CrossRefGoogle Scholar
  11. 11.
    Goerick C, Mikhailova I, Wersing H, Kirstein S (2006) Biologically motivated visual behaviours for humanoids: learning to interact and learning in interaction. In: Proceedings of the IEEE/RSJ international conference on humanoid robotsGoogle Scholar
  12. 12.
    Guyon I, Elissee A (2003) An introduction to variable and feature selection. J Mach Learn Res 3: 1157–1182zbMATHCrossRefGoogle Scholar
  13. 13.
    Hamker FH (2001) Life-long learning cell structures—continously learning without catastrophic interference. Neural Netw 14: 551–573CrossRefGoogle Scholar
  14. 14.
    Hammer B, Villmann T (2002) Generalized relevance learning vector quantization. Neural Netw 15(8–9): 1059–1068CrossRefGoogle Scholar
  15. 15.
    Hasler S, Wersing H, Körner E (2007) A comparison of features in parts-based object recognition hierarchies. In: Proceedings of the international conference on artificial neural networks (ICANN), pp 210–219Google Scholar
  16. 16.
    Kirstein S, Wersing H, Körner E (2008) A biologically motivated visual memory architecture for online learning of objects. Neural Netw 21: 65–77Google Scholar
  17. 17.
    Kirstein S, Wersing H, Gross H-M, Körner E (2008) A vector quantization approach for life-long learning of categories. In: Proceedings international conference on neural information processing (ICONIP). Springer, pp 803–810Google Scholar
  18. 18.
    Kohonen T (1989) Self-organization and associative memory. Springer Series in information sciences, 3rd edn. SpringerGoogle Scholar
  19. 19.
    Leibe B, Leonardis A, Schiele B (2004) Combined object categorization and segmentation with an implicit shape model. In: In ECCV workshop on statistical learning in computer vision, pp 17–32Google Scholar
  20. 20.
    Lowe DG (2004) Distinctive image features from scale-invariant keypoints. Int J Comp Vis 60(2): 91–110CrossRefGoogle Scholar
  21. 21.
    Mikolajczyk K, Leibe B, Schiele B (2006) Multiple object class detection with a generative model. In: Proceedings IEEE conference on computer vision and pattern recognition (CVPR)Google Scholar
  22. 22.
    Ozawa S, Toh SL, Abe S, Pang S, Kasabov N (2005) Incremental learning of feature space and classifier for face recognition. Neural Netw 18(5–6): 575–584CrossRefGoogle Scholar
  23. 23.
    Pomierski T, Gross HM (1996) Biological neural architecture for chromatic adaptation resulting in constant color sensations. In: Proceedings IEEE international conference on neural networks (ICNN), pp 734–739Google Scholar
  24. 24.
    Roth PM, Donoser M, Bischof H (2006) On-line learning of unknown hand held objects via tracking. In: Proceedings of the second international cognitive vision workshop (ICVW)Google Scholar
  25. 25.
    Schneider P, Biehl M, Hammer B (2007) Relevance matrices in LVQ. In: Similarity-based clustering and its application to medicine and biology, Number 07131 in Dagstuhl seminar proceedings. Internationales Begegnungs- und Forschungszentrum für Informatik (IBFI), Schloss Dagstuhl, GermanyGoogle Scholar
  26. 26.
    Skočaj D, Berginc G, Ridge B, Štimec A, Jogan M, Vanek O, Leonardis A, Hutter M, Hewes N (2007) A system for continuous learning of visual concepts. In: Proceedings of the international conferance on vision systems (ICVS)Google Scholar
  27. 27.
    Skočaj D, Kristan M, Leonardis A (2008, January) Continuous learning of simple visual concepts using incremental kernel density estimation. In: Proceedings of the international conference on computer vision theory and applications (VISAPP), Funchal, Madeira, Portugal, pp 598–604Google Scholar
  28. 28.
    Steels L, Kaplan F (2001) AIBO’s first words. The social learning of language and meaning. Evolut Commun 4(1): 3–32CrossRefGoogle Scholar
  29. 29.
    Swain MJ, Ballard DH (1991) Color indexing. Int J Comput Vis 7(1): 11–32CrossRefGoogle Scholar
  30. 30.
    Thomas A, Ferrari V, Leibe B, Tuytelaars T, Schiele B, Gool LV (2006, June). Towards multi-view object class detection. In: Proceedings IEEE conference on computer vision and pattern recognition (CVPR), New York, USAGoogle Scholar
  31. 31.
    Wersing H, Körner E (2003) Learning optimized features for hierarchical models of invariant object recognition. Neural Comput 15(7): 1559–1588zbMATHCrossRefGoogle Scholar
  32. 32.
    Wersing H, Kirstein S, Götting M, Brandl H, Dunn M, Mikhailova I, Goerick C, Steil J, Ritter H, Körner E (2007) Online learning of objects in a biologically motivated architecture. Int J Neural Sys 17: 219–230CrossRefGoogle Scholar
  33. 33.
    Willamowski J, Arregui D, Csurka G, Dance CR, Fan L (2004) Categorizing nine visual classes using local appearance descriptors. In: Proceedings of the ICPR workshop on learning for adaptable visual systemsGoogle Scholar

Copyright information

© Springer-Verlag 2009

Authors and Affiliations

  • Stephan Kirstein
    • 1
    • 2
  • Alexander Denecke
    • 1
    • 3
  • Stephan Hasler
    • 1
  • Heiko Wersing
    • 1
  • Horst-Michael Gross
    • 2
  • Edgar Körner
    • 1
  1. 1.Honda Research Institute Europe GmbHOffenbach am MainGermany
  2. 2.Ilmenau University of Technology, Neuroinformatics and Cognitive Robotics LabIlmenauGermany
  3. 3.Bielefeld University, CoR-LabBielefeldGermany

Personalised recommendations