Title :
Robust CNN templates: theory and simulations
Author :
Mirzai, Bahram ; Lím, Drahoslalv ; MOSCHYT, GEORGE S.
Author_Institution :
Signal & Inf. Process. Lab., Swiss Federal Inst. of Technol., Zurich, Switzerland
Abstract :
In an analog CNN circuit, the accuracy of the template and data values will always suffer from various kinds of errors. How large these errors can be before they jeopardize correct CNN operation, is an issue important for the design of CNN circuits. We present design approaches applicable to several CNN processing tasks, which guarantee the correct operation of the network even when the template and data values are subjected to some random variations. The robustness of the proposed templates is verified by simulations
Keywords :
analogue integrated circuits; cellular neural nets; edge detection; integrated circuit design; neural chips; analog CNN circuit; design approaches; robust CNN templates; Cellular neural networks; Circuit simulation; Computer errors; Error correction; Information processing; Laboratories; Nonlinear systems; Robustness; Signal processing; Very large scale integration;
Conference_Titel :
Cellular Neural Networks and their Applications, 1996. CNNA-96. Proceedings., 1996 Fourth IEEE International Workshop on
Conference_Location :
Seville
Print_ISBN :
0-7803-3261-X
DOI :
10.1109/CNNA.1996.566606
