Title :
Neural network based universal tone detector
Author :
Sträußnigg, D. ; Schwingshackl, D. ; Schaller, M.
Abstract :
This paper presents a new algorithm for universal tone detection, which uses two outputs of the Goertzel filter as inputs to a neural network. This algorithm is referred to as the neural tone detector (NTD). Detailed analysis and comparison with existing solutions in terms of performance and computational complexity is given. The NTD was found to be a reliable high performance solution for the tone detector under noisy conditions - whereas computational complexity and memory requirements are low.
Keywords :
computational complexity; neural nets; signal detection; Goertzel filter; computational complexity; low memory requirements; neural network; neural tone detector; noisy conditions; universal tone detector; Artificial neural networks; Computational complexity; Detection algorithms; Detectors; Filters; Frequency; Neural networks; Robustness; Signal processing algorithms; Working environment noise;
Conference_Titel :
Circuits and Systems, 2007. MWSCAS 2007. 50th Midwest Symposium on
Conference_Location :
Montreal, Que.
Print_ISBN :
978-1-4244-1175-7
Electronic_ISBN :
1548-3746
DOI :
10.1109/MWSCAS.2007.4488603
