Title :
What to do first: The initial behavior in a multi-sensory household object recognition and categorization system
Author :
Haojun Guan ; Weipeng He ; Jianwei Zhang
Author_Institution :
Dept. of Inf., TAMS Group, Univ. of Hamburg, Hamburg, Germany
Abstract :
Numbers of studies in Biological area have already shown that our human beings have the amazing ability to quickly extract the physical properties of an object from the sound it made and recognize at least the material of it [1][2]. The interactive exploratory behaviours are used as a “question” which “ask” to the object from human, and the feedback, maybe a feeling or a sound, as the “answer”. Such an intelligent system can also be developed based on the research above, the ability of recognize and category objects is also necessary for the intelligent system and robot especially service robot in the near future. In this paper, a novel framework of a multi-sensory household object recognition and categorization system is introduced. Both auditory and visual sensory are used in this novel system, the detail of the part of auditory is presented in detail, and the result of the experiments to decide which interactive behaviour is most suitable act as the initial behaviour is also shown.
Keywords :
intelligent robots; object recognition; sensor fusion; service robots; auditory sensory; categorization system; category object; intelligent system; interactive behaviour; interactive exploratory behaviour; multisensory household object recognition; service robot; visual sensory; Accuracy; Bismuth; Machine learning algorithms; Silicon; Spectrogram; Support vector machines;
Conference_Titel :
Multisensor Fusion and Information Integration for Intelligent Systems (MFI), 2014 International Conference on
Conference_Location :
Beijing
Print_ISBN :
978-1-4799-6731-5
DOI :
10.1109/MFI.2014.6997736