Title :
A time-frequency neutral network layered model for hearing perception
Author :
Georgopoulos, Voula C. ; Preis, Douglas
Author_Institution :
Dept. of Electr. Eng., Tufts Univ., Medford, MA, USA
Abstract :
This paper introduces a layered neural network model for hearing perception. It is based on five important perceptual properties of hearing. The neural network model processes a joint-domain representation of the input signal to yield the desired perceptual properties. The focus is on the first two layers of the model, the transformation layer and two feature extraction layers
Keywords :
acoustic signal processing; feature extraction; feedforward neural nets; hearing; multilayer perceptrons; time-frequency analysis; feature extraction layers; hearing perception; input signal; joint-domain representation; layered neural network model; perceptual properties; time-frequency analysis; transformation layer; Auditory system; Biological neural networks; Biological system modeling; Biological systems; Biology computing; Frequency conversion; Neurons; Psychology; Signal processing; Time frequency analysis;
Conference_Titel :
Applications of Signal Processing to Audio and Acoustics, 1993. Final Program and Paper Summaries., 1993 IEEE Workshop on
Conference_Location :
New Paltz, NY
Print_ISBN :
0-7803-2078-6
DOI :
10.1109/ASPAA.1993.379981
