Abstract:
Vital signal renovation plays an important role in a wide range of applications, including signal analysis and diagnosing diseases through it. Therefore, it is salient to get the main content of the vital signal. In this research, a new approach to the problem of noise removal from vital signals is presented based on random optimization through Monte Carlo Markov Chain (MCMC) sampling. For this purpose, the problem of noise omission from the vital signal is described as a Bayesian squared minimization problem, and considering a non-parametric random approach to solve this problem, the Monte Carlo Markov Chain noise omission approach is flexibly adapted to the noise detection domain in vital signals. To test the performance of the proposed method, four types of vital signals have been used: Medical images, ECG electrocardiogram signals, EEG brain signals as well as ENG nerve and muscle signals. The results of the experiments show that the use of sampling technique based on Gaussian distribution and, retrieving the signal based on the weighted average in the selected samples allows a more accurate estimate of the ideal signal. This more accurate estimation minimizes the difference between the actual and the retrieved signals. As a result, in addition to reducing the mean error squares, the signal-to-noise ratio increases.
摘要:
生命信号改造在广泛应用中发挥着重要作用,包括信号分析和通过它诊断疾病。因此,重要的是获取生命信号的主要内容。在这项研究中,通过蒙特卡罗马尔可夫链(MCMC)采样,提出了一种基于随机优化的从生命信号中去除噪声的新方法。为此,生命信号中的噪声遗漏问题被描述为贝叶斯平方最小化问题,并考虑一种非参数随机方法来解决这个问题,蒙特卡洛马尔可夫链噪声遗漏方法灵活地适应于生命信号中的噪声检测域。为了测试所提出的方法的性能,已经使用了四种类型的生命信号:医学图像,心电图心电图信号,EEG脑信号以及ENG神经和肌肉信号。实验结果表明,采用基于高斯分布的采样技术,基于所选样本中的加权平均值检索信号允许对理想信号的更准确估计。这种更准确的估计使实际信号和检索信号之间的差异最小化。因此,除了减少平均误差平方,信噪比增加。