DocumentCode :
1731048
Title :
Accurate CMOS implementation of PWL CNN neuron activations
Author :
Forti, M. ; Pancioni, L. ; Rocchi, S. ; Vignoli, V.
Author_Institution :
Dept. of Inf. Eng., Siena Univ., Italy
Volume :
1
fYear :
2002
fDate :
6/24/1905 12:00:00 AM
Abstract :
The paper proposes a new electronic design which closely approximates the standard implementation in CMOS technology for the neuron cell of the Cellular Neural Networks (CNNs) introduced by Chua and Yang (see IEEE Trans. Circuits Syst., vol. 35, p. 1257-1272, 1988). This permits us to obtain an input-output neuron activation ideal piecewise linear (PWL) activation with three segments. The possibility of achieving such a good approximation is of great interest in that it makes it possible to exploit the huge body of efficient design techniques for PWL CNNs already available in the literature.
Keywords :
CMOS analogue integrated circuits; analogue processing circuits; cellular neural nets; neural chips; piecewise linear techniques; CMOS implementation; PWL CNN neuron activations; cellular neural networks; input-output neuron activation; neuron cell; piecewise linear activation; CMOS technology; Cellular neural networks; Circuits; Differential amplifiers; Hopfield neural networks; Mirrors; Neural networks; Neurons; Piecewise linear approximation; Piecewise linear techniques;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Circuits and Systems, 2002. ISCAS 2002. IEEE International Symposium on
Print_ISBN :
0-7803-7448-7
Type :
conf
DOI :
10.1109/ISCAS.2002.1009817
Filename :
1009817
Link To Document :
بازگشت