Share Email Print

Proceedings Paper

Fast implementation of neural network classification
Author(s): Guiwon Seo; Jiheon Ok; Chulhee Lee
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Most artificial neural networks use a nonlinear activation function that includes sigmoid and hyperbolic tangent functions. Most artificial networks employ nonlinear functions such as these sigmoid and hyperbolic tangent functions, which incur high complexity costs, particularly during hardware implementation. In this paper, we propose new polynomial approximation methods for nonlinear activation functions that can substantially reduce complexity without sacrificing performance. The proposed approximation methods were applied to pattern classification problems. Experimental results show that the processing time was reduced by up to 50% without any performance degradations in terms of computer simulation.

Paper Details

Date Published: 24 September 2013
PDF: 8 pages
Proc. SPIE 8871, Satellite Data Compression, Communications, and Processing IX, 887107 (24 September 2013); doi: 10.1117/12.2026666
Show Author Affiliations
Guiwon Seo, Yonsei Univ. (Korea, Republic of)
Jiheon Ok, Yonsei Univ. (Korea, Republic of)
Chulhee Lee, Yonsei Univ. (Korea, Republic of)

Published in SPIE Proceedings Vol. 8871:
Satellite Data Compression, Communications, and Processing IX
Bormin Huang; Antonio J. Plaza; Chein-I Chang, Editor(s)

© SPIE. Terms of Use
Back to Top