Abstract:
This paper presents a serial hardware architecture of multilayer perceptron (MLP) neural network to classify respiratory sounds into normal and wheezing classes. Whatever...Show MoreMetadata
Abstract:
This paper presents a serial hardware architecture of multilayer perceptron (MLP) neural network to classify respiratory sounds into normal and wheezing classes. Whatever the number of neurons in the MLP network, this architecture uses a single multiply-accumulate (MAC) block, two memory blocks, and a control block based on binary counters. It has been efficiently implemented on FPGA chip using a high abstraction level programming tool, Xilinx System Generator (XSG) in MATLAB/SIMULINK environment. The serial architecture gives the same classification performances as the parallel one, but it presents the advantage of using much less hardware resources.
Date of Conference: 09-11 August 2021
Date Added to IEEE Xplore: 13 September 2021
ISBN Information: