VFA-driven Hierarchical Temporal Memory input for object categorization

Csapó Ádám, Péter Baranyi

Research output: Contribution to conferencePaper

1 Citation (Scopus)

Abstract

Results in visual psychology have shown that the location and statistics of nodes, endpoints and corners carry essential and sufficient information for object recognition. In this paper, we present a method for object categorization which relies on the combination of the Visual Feature Array model and Hierarchical Temporal Memories. Experimental results show that even without taking into consideration statistics other than the spatial distribution of nodes, two categories can be told apart with a success rate of about 95%. The same results could not be achieved by simply taking into account grayscale pixel values. Efforts were also made to generalize the above results to a categorization task among 10 different categories.

Original languageEnglish
Publication statusPublished - Dec 1 2007
Event8th International Symposium of Hungarian Researchers on Computational Intelligence and Informatics, CINTI 2007 - Budapest, Hungary
Duration: Nov 15 2007Nov 17 2007

Other

Other8th International Symposium of Hungarian Researchers on Computational Intelligence and Informatics, CINTI 2007
CountryHungary
CityBudapest
Period11/15/0711/17/07

ASJC Scopus subject areas

  • Artificial Intelligence
  • Information Systems

Fingerprint Dive into the research topics of 'VFA-driven Hierarchical Temporal Memory input for object categorization'. Together they form a unique fingerprint.

  • Cite this

    Ádám, C., & Baranyi, P. (2007). VFA-driven Hierarchical Temporal Memory input for object categorization. Paper presented at 8th International Symposium of Hungarian Researchers on Computational Intelligence and Informatics, CINTI 2007, Budapest, Hungary.