A Segmentation Method for Noisy Speech Using Genetic Algorithm

Abstract

This paper presents a technique to automatically segment a speech signal in noisy environments. The speech segmentation is formulated as an optimization problem and boundaries of the speech segments are detected using genetic algorithm (GA). The initial number of segments is estimated from the modified version of the signal using the minimal number of binary Walsh basis functions. The segmentation results are improved through the generations of GA by introducing a new evaluation function, which is based on the sample entropy and a heterogeneity measure. The results of the experiments, which have been carried out on TIDIGITS database and different types and levels of noise, show the efficiency of the proposed genetic segmentation algorithm.

The pairwise approximate spatiotemporal symmetry algorithm: A method for segmenting time series pairs.

“…The problem of multivariate time-series segmentation has been approached from a number of directions. Omranian et al (2015) suggested four broad categories of segmentation approaches which are particularly relevant to biological data: clustering (Abonyi et al, 2005; Maya et al, 2020), graphical models (Angelosante & Giannakis, 2011; Xuan & Murphy, 2007), genetic algorithms (Nikolaou et al, 2015; Pwint & Sattar, 2005), and regression (Chamroukhi et al, 2013; Omranian et al, 2015). These modeling approaches are parametric in nature.…”

An evolutionary approach for segmentation of noisy speech signals for efficient voice activity detection

“…The experimental results show that the proposed method can detect the boundaries of speech and non-speech events with high accuracy in various noise conditions, based on work described in Ref. [16] 2. RELATED WORK Sohn et al(1999) [17] proposed a voice activity detector (VAD) based on statistical models, which employ the decision-directed parameter estimation method for the likelihood ratio test together with a hang-over scheme using HMMs.…”

Bio inspired Cuckoo Search Algorithm based neural network and its application to noise cancellation

“…In [6] a fuzzy logic based neural network has been employed for adaptive noise cancellation Evolutionary techniques can also be used for noise removal. In [7], Genetic Algorithm has been used for this purpose. In this paper Cuckoo Search algorithm, based on the brooding characteristics of Cuckoo birds is used for training the neural network.…”