DocumentCode :
264059
Title :
Design space exploration for a single-FPGA handwritten digit recognition system
Author :
Thang Viet Huynh
Author_Institution :
Danang Univ. of Sci. & Technol., Danang, Vietnam
fYear :
2014
fDate :
July 30 2014-Aug. 1 2014
Firstpage :
291
Lastpage :
296
Abstract :
Multilayer perceptron neural networks have widely been implemented on reconfigurable hardware to perform a variety of applications including classification and pattern recognition. This paper investigates the combined impact of neural network size and reduced precision number formats, used for the representation of the optimal parameters, on the recognition rate a neural network based handwritten digit recognition system. The MNIST database is used for training and testing in this work. After deriving the optimal reduced-precision floating-point format sufficient for achieving a desired recognition performance, we provide an estimate for the hardware resources needed to implement the network on FPGAs. Our work allows for an efficient investigation of tradeoffs in operand word-length, network size, recognition rate and hardware cost of reduced-precision neural network implementations on reconfigurable hardware.
Keywords :
field programmable gate arrays; floating point arithmetic; handwritten character recognition; multilayer perceptrons; MNIST database; design space exploration; multilayer perceptron neural networks; network size; operand word-length; pattern classification; pattern recognition; precision number formats reduction; recognition rate; reconfigurable hardware; single-FPGA handwritten digit recognition system; Adders; Field programmable gate arrays; Handwriting recognition; Neurons; Standards; FloPoCo; MNIST; MPFR; bit width allocation; floating-point; neural network; reconfigurable hardware;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Communications and Electronics (ICCE), 2014 IEEE Fifth International Conference on
Conference_Location :
Danang
Print_ISBN :
978-1-4799-5049-2
Type :
conf
DOI :
10.1109/CCE.2014.6916717
Filename :
6916717
Link To Document :
بازگشت