VSE knjižnice (vzajemna bibliografsko-kataložna baza podatkov COBIB.SI)
  • Invariant pattern classification neural network versus FT approach
    Dobnikar, Andrej ...
    Invariant Pattern Classification (IPC) of black and white images is studied in two directions. Classical Fourier Transform (FT) approach consists of normalisation procedure for excluding ... magnification or size influence and FT transforms, whose output spectrum descriptors are invariant to rotation and/or translation. Only pattern calssification independent of size and rotation is concerned in this paper. It is supposed that the observed object is placed in the center of the image. An alternative approach uses Neural Network Modeling (NNM) to achieve the same effect. It is shown that the proposed Neural Network with the combination of dynamic and associative properties better solves the same problem with additional potential capability of considerable higher parallel processing.
    Vir: Microprocessing and microprogramming. - ISSN 0165-6074 (Vol. 33, no. 3, 1992, str. 161-168)
    Vrsta gradiva - članek, sestavni del
    Leto - 1992
    Jezik - angleški
    COBISS.SI-ID - 1806187

Vnos na polico