基于双树复小波变换的脑血栓检测研究
Detection of Cerebral Thrombosis Based on Dual Tree Complex Wavelet Transform

DOI: 10.12677/HJBM.2019.92008, PP. 49-56

Keywords: 双树复小波变换，降维，支持向量机，脑血栓
Dual-Tree Complex Wavelet Transform, Dimen-sionality Reduction, Support Vector Machines, Cerebral Thrombosis

Full-Text   Cite this paper   Add to My Lib

Abstract:

脑动脉中的栓子有可能阻塞脑血管而引起缺血性脑卒中等疾病，因此脑血栓检测具有重要的临床意义。传统的脑血栓检测依赖于专家的主观判断，并且耗时长。为了克服这些缺点，本文提出了一种基于双树复小波变换(DTCWT)的快速、准确和健壮的脑血栓检测方法。双树复小波变换相对普通离散小波变换，增强了从多普勒超声信号中提取的系数的鲁棒性。对于本文提出的检测方法，采用栓塞和人工伪像信号进行实验论证，首先对不同样本的正向血流信号使用DTCWT分别提取其血流系数，然后对每组系数进行降维，并将降维后的系数逐一馈送进分类器。将所得结果与基于FFT和DWT的脑血栓检测方法进行比对，结果表明使用DTCWT提取的特征具有最高的准确性和栓塞检测率。
The embolus in the cerebral arteries may block the cerebral blood vessels and cause ischemic stroke. Therefore, cerebral thrombosis detection has important clinical significance. Traditional cerebral thrombosis testing relies on subjective judgment by experts and is time consuming. In order to overcome these shortcomings, this paper proposes a fast, accurate and robust method for detecting cerebral thrombosis based on dual-tree complex wavelet transform (DTCWT). The dual-tree complex wavelet transform is more robust than the ordinary discrete wavelet transform, which enhances the coefficients extracted from the Doppler ultrasound signal. For the detection method proposed in this paper, embolization and artificial artifact signals are used for experimental demonstration. Firstly, the blood flow coefficients of the forward blood flow signals of different samples are extracted by DTCWT, then the dimensionality of each set of coefficients is reduced and the dimension is reduced. The subsequent coefficients are fed into the classifier one by one. Comparing the results with the FFT and DWT-based cerebral thrombus detection system, the results showed that the features extracted using DTCWT had the highest accuracy and embolization rate.

References

[1]	Markus, H.S. (2000) Monitoring Embolism in Real Time. Circulation, 102, 826-828. https://doi.org/10.1161/01.CIR.102.8.826
[2]	King, A., Shipley, M. and Marksu, H. (2012) Optimizing Protocols for Risk Prediction Asymptomatic Carotid Stenosis Using Embolic Signal Detection: The Asymptomatic Carotid Emboli Study. Journal of Vascular Surgery, 55, 1-1220. https://doi.org/10.1016/j.jvs.2012.02.016
[3]	Mitsuoka, A., Inoue, Y., Kume, H., et al. (2010) Discrimination of Types of Venous Emboli Using Doppler Ultrasound. Annals of Vascular Surgery, 24, 721-727. https://doi.org/10.1016/j.avsg.2010.03.004
[4]	Guepie, B.K., Sciolla, B., Millioz, F., et al. (2017) Discrimination between Emboli and Artifacts for Outpatient Transcranial Doppler Ultrasound Data. Medical & Biological Engineering & Computing, 55, 1-11.
[5]	Aydin, N. (2007) DWT Based Adaptive Threshold Determination in Embolic Signal Detection. Nasa/Esa Conference on Adaptive Hardware & Systems, Edinburgh, 5-8 August 2007, 214-219.
[6]	Xu, D. and Wang, Y. (2007) An Automated Feature Extraction and Emboli Detectionsystem Based on the PCA and Fuzzy Sets. Computers in Biology and Medicine, 37, 861-871.
[7]	Baudry, M. (1998) The Narrow Band Hypothesis: An Interesting Approach for High Intensity Transient Signals (HITS) Detection. Ultrasound in Medicine & Biology, 24, 375. https://doi.org/10.1016/S0301-5629(97)00279-2
[8]	Serbes, G. and Aydin, N. (2011) Denoising Embolic Doppler Signals Using Modified Complex Discrete Wavelet Transform. IEEE Signal Processing & Communications Applications Conference, Antalya, 20-22 April 2011, 1840-1843.
[9]	Aydin, N., Marvasti, F. and Markus, H.S. (2004) Embolic Doppler Ultrasound Signal Detection Using Discrete Wavelet Transform. IEEE Transactions on Information Technology in Biomedicine, 8, 182-190. https://doi.org/10.1109/TITB.2004.828882
[10]	Selesnick, I.W., Baraniuk, R.G. and Kingsbury, N.C. (2005) The Dual-Tree Complex Wavelet Transform. IEEE Signal Processing Magazine, 22, 123-151. https://doi.org/10.1109/MSP.2005.1550194
[11]	Serbes, G., Sakar, C.O., Kahya, Y., et al. (2012) Pulmonary Crackle Detection Using Time-Frequency Analysis. Signal Processing & Communications Applications Conference, Mugla, 18-20 April 2012, 1012-1021.
[12]	Kingsbury, N.G. (1999) Image Processing with Complex Wavelets. Phil-osophical Transactions of the Royal Society, 357, 2543-2560. https://doi.org/10.1098/rsta.1999.0447
[13]	Yang, J. and Yang, J.Y. (2003) Why Can LDA Be Performed in PCA Transformed Space. Pattern Recognition, 36, 563-566. https://doi.org/10.1016/S0031-3203(02)00048-1
[14]	Cherkassky, V. and Ma, Y. (2004) Practical Selection of SVM Parameters and Noise Estimation for SVM Regression. Neural Networks, 17, 113-126. https://doi.org/10.1016/S0893-6080(03)00169-2
[15]	Bao, Z.J., Zhong, W.M., Pi, D.Y., et al. (2007) Robustly Stable Model Predictive Control Based on Parallel Support Vector Machines with Linear Kernel. Journal of Central South University of Technology, 14, 701-707. https://doi.org/10.1007/s11771-007-0134-9

Full-Text