Intelligent detection of DTMF tones using a hybrid signal processing technique with support vector machines

Abstract

Efficient methods for DTMF signal detection are important for developing telecommunication equipment. This paper presents a hybrid signal processing and artificial intelligence based approach for the detection of Dual-tone Multifrequency (DTMF) tones under the influence of White Gaussian Noise (WGN) and frequency variation. Key innovations include the use of a Finite Impulse Response (FIR) bandpass filter for reduction of noise from DTMF input samples, and Support Vector Machines (SVM) for intelligent classification of the detected DTMF carrier frequencies. The proposed hybrid DTMF detector scheme is based on power spectrum analysis by means of the Discrete Fourier Transform (DFT). The Goertzel's Algorithm is used to estimate the seven fundamental DTMF carrier frequencies. The tone detection scheme employs decision logic to detect valid DTMF tones from low and high DTMF frequency groups. Comparison of this hybrid DTMF tone detection model with existing DTMF detection techniques proves the merits of this proposed scheme.