Adaptive filtering noisy transcranial Doppler signal by using artificial bee colony algorithm

https://doi.org/10.1016/j.engappai.2012.10.011Get rights and content

Abstract

Computerized processes are supportive in the new age of medical treatment. Biomedical signals which are collected from the human body supply or important useful data that are related with the biological actions of human body organs. However, these signals may also contain some noise. Heart waves are commonly classified as biomedical signals and are non-stationary due to their statistical specifications. The probability distributions of the noise are very different, and for this reason there is no common method to remove the noise. In this study, adaptive filters are used for noise elimination and the transcranial Doppler signal is analyzed. The artificial bee colony algorithm was employed to design the adaptive IIR filters for noise elimination on the transcranial Doppler signal and the results were compared to those obtained by the methods based on popular and recently introduced evolutionary algorithms and conventional methods.

Introduction

The Doppler ultrasound device provides non-invasive measurement of blood flow velocity with the aim of diagnosing vascular diseases. The Doppler shifts from red blood cells are used for computing blood velocity. In the computing of blood velocity, the Doppler-shift frequency and Doppler angle are used by the instrument. The Doppler signal displays a time-varying random character because the signal back scattered from the blood possesses a random spatial distribution (Evans, 1989). Doppler indices such as the resistance index (RI) or pulsatility index (PI) are ratios that are computed from various points on the spectrum which are computed to analyze the Doppler signals (Diniz, 2008). Traditionally, the short time Fourier transform (STFT) method has been found to be suitable for the application of the spectrogram (Behbahani, 2007, Billings and Fung, 1995, Brody and Meindl, 1974, Evans, 1989). Doppler spectrogram indices are determined from the maximum frequency waveform of the Doppler spectrogram (Diniz, 2008, Feder et al., 1989). The estimation resolution of the maximum frequency waveform is affected by the inner or outer noise in the system causing extra frequency. Hence, this is a significant stage which includes the denoising of the Doppler ultrasound signal for further processing (Haykin, 1996, Haykin, 2002).

When the system parameters or signal conditions change, adaptive filters are generally used and they are to be adjusted to balance this change (Behbahani, 2007). It is known that all adaptive filters capable of adapting at real-time rates experience losses in performance because their adjustments are based on statistical averages taken with limited sample sizes (Widrow, 1971). In the adaptive filtering case, the parameters of the filter which were evaluated a few moments before are used to automatically tune the parameters of the filter which are determined at the present moment, to adapt to the changing situation due to the achievement of the optimal filtering (He et al., 2008). The adaptive filter has the most important properties, because it can be effectively applied in unpredictable situations and the input signal the characteristics of which vary with time, may be tracked by using it. The adaptive filter has been applied mainly in signal processing, control, communications and many other systems for noise cancellation. The gradient based algorithms move in the negative gradient direction. So, they aim to obtain the global minimum on the error surface. However, the filter design approaches which are based on the gradient algorithms may lead to suboptimal IIR filter designs when the error surface is multi-modal (Karaboga and Basturk, 2007, Karaboga and Basturk, 2008, Karaboga, 2005, Karaboga et al., 2012, Karaboga, 2005).

The artificial bee colony (ABC) algorithm is a relatively new swarm intelligence based algorithm which can be used to find optimal or near optimal solutions to numerical and discrete problems. The ABC algorithm was first introduced by Karaboga in 2005, inspired by the foraging behavior of honeybees (Karaboga, 2009, Karaboga, 2005). It is simple and robust optimization algorithm which can be easily implemented in widely used programming languages and has proven to be both very effective and quick for a diverse set of optimization problems (Karaboga et al., in press). Because of the nonstationary character of Doppler signals, the practical issues related to this signal must be solved using adaptive filters (Kaluzynski, 1987). In this work, a novel approach based on the ABC algorithm is proposed to denoise the Doppler signal by using adaptive IIR filter structures and also, its performance is compared to the popular algorithms such as particle swarm optimization (PSO) and differential evolution (DE), and conventional wavelet transform techniques. The paper is organized as follows. Section 2 describes the proposed artificial bee colony based adaptive noise cancellation system. Section 3 presents the application of the proposed method to the noise cancellation problem. The simulation study is outlined in Section 4 and the simulation results are discussed in Section 5.

Section snippets

Proposed adaptive noise cancellation system

Widrow and Glover proposed adaptive noise cancellation in 1975 (Widrow and Glover, 1975). Noise cancellation is a key problem of filter design. This filter design can be applied when the reference noise signal is achieved. In many applications, e.g. speech processing, echo cancellation and enhancement, antenna array processing, biomedical signal and image processing, the noise cancellation technique has been used (Widrow and Walach, 1996, Liu et al., 1997, Liu et al., 1999, Ng et al., 1996). In

Artificial bee colony algorithm

Swarm intelligence is a popular area in the field of optimization and researchers have developed various algorithms by modeling the behaviors of different swarms of animals and insects such as ants, termites, bees, birds and fish (Karaboga and Akay, 2009). Some classical kinds of swarms include a colony of ants; a bee colony swarming around their hive; a swarm of cells is an immune system, a crowd which is a swarm of people; and a flock of birds.

The artificial bee colony algorithm was

Simulation study

In this study, simulated Doppler signals were obtained using the method introduced by Wang and Fish (Wang and Fish, 1996). The proposed method supposes that the Doppler signal is a Gaussian random process. If a filter is linear and it is stimulated by additive white Gaussian noise (AWGN), the output of that filter will be a Gaussian random process. That is why the Doppler signal from the transcranial arteries can be generated by AWGN input into a time-varying filter. This filter has a h(j,n)

Results and discussion

The study was performed using both the recorded and simulated transcranial Doppler (TCD) signal. The clinical TCD signals were sampled at 44,100 Hz with a linear ultrasound probe of 5 MHz. These samples were fed into the adaptive noise canceler used. The corrupted TCD signal was the primary input into the filter. The adaptive filter was implemented as IIR structures with one and five coefficients. In other words, the order of filters is 1 and 5. The coefficients were adjusted by using the ABC

Conclusion

An adaptive IIR filter design method based on ABC algorithm was described for the elimination of noise on the transcranial Doppler signal. The performance of the proposed method was also compared to the conventional wavelet transform based methods and popular evolutionary algorithms. The simulation results obtained showed that an adaptive IIR filter can be efficiently designed by the proposed method using a window size of 1000 samples for denoising the disturbed transcranial Doppler signal.

References (30)

  • M. Feder et al.

    Maximum likelihood noise cancellation using the EM algorithm

    IEEE Trans. Acoust. Speech Signal Process.

    (1989)
  • S. Haykin

    Adaptive Filter Theory

    (1996)
  • S. Haykin

    Adaptive Filter Theory

    (2002)
  • He, Y., He, H., Li, L., Wu, Y., Pan, H., 2008. The applications and simulation of adaptive filter in noise canceling....
  • Karaboga, D., 2005. An Idea Based on Honey Hee Swarm for Numerical Optimization. Technical Report-TR06 . Engineering...
  • Cited by (33)

    • Dynamic takt time decisions for paced assembly lines balancing and sequencing considering highly mixed-model production: An improved artificial bee colony optimization approach

      2021, Computers and Industrial Engineering
      Citation Excerpt :

      Through metaheuristic random search and greedy selection, the ABC algorithm can cause the entire population to evolve to obtain the optimal solution (Karaboga & Basturk, 2007; Karaboga & Ozturk, 2011). The standard ABC algorithm has been widely used in many fields (Karaboga and Latifoglu, 2013A; Karaboga and Latifoglu, 2013B; Yildiz, 2013). Although the standard ABC algorithm offers many advantages, it cannot be directly used to solve multi-objective optimization problems.

    • Classification of transcranial Doppler signals using individual and ensemble recurrent neural networks

      2017, Neurocomputing
      Citation Excerpt :

      In [5], Doppler techniques were deployed for measuring the blood flow velocity in diagnosing vascular diseases. Specifically, shifts from the red blood cells detected by the Doppler technique were used for computing the blood velocity [5]. Indeed, Doppler sonography is typically preferred since angiography is costly and is an invasive method [6].

    • Wavenet using artificial bee colony applied to modeling of truck engine powertrain components

      2015, Engineering Applications of Artificial Intelligence
      Citation Excerpt :

      Some attempts have been made to develop nonlinear identification approaches to automotive applications. In this context, see examples in (Keen and Cole, 2012; Ye, 2007; Liu and Bewley, 2003; Zito and Landau, 2005; Wu and Liu, 2009; Worden et al., 2001; Tan and Saif, 2000; Togun et al., 2012; Antory, 2007; Coelho et al., 2014). Recently, natural computing techniques, such as artificial neural networks (ANNs) (Kilic et al., 2014; Gil et al., 2013), fuzzy systems (Babuška and Verbruggen, 1996; Dovžan and Škrjanc, 2010) and combined neuro-fuzzy systems (Salahshoor et al., 2012) have become effective tools of identification of nonlinear processes, especially to those systems whose mathematical models are difficult to obtain.

    • Two modified Artificial Bee Colony algorithms inspired by Grenade Explosion Method

      2015, Neurocomputing
      Citation Excerpt :

      Their results [2,4–11] demonstrate that ABC is simple in concept, few in parameters, easy for implementation and more effective than some other population-based algorithms such as Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Ant Colony Optimization (ACO) and Differential Evolution (DE) algorithm. Therefore, ABC has aroused much interest and has been successfully used in different fields, including neural networks training [5,12,13], data clustering [7,14,15], classification [16,17], financial forecasting [18], signal processing [19,20], image processing [21], pattern recognition [22], scheduling [3,23], control system design [24], network communication [25], identification of structural model [26], and other optimizations [1,2,4,6,8–11,27–30]. Exploration and exploitation are extremely important mechanisms in a robust search process.

    • A novel 2D-ABC adaptive filter algorithm: A comparative study

      2015, Digital Signal Processing: A Review Journal
    View all citing articles on Scopus
    View full text