Fast denoising of multi-channel transcranial magnetic stimulation signal based on improved generalized mathematical morphological filtering. (February 2022)

• Record Type: Journal Article
• Authors: Liu, Jinzhen; Tian, Kaiwen; Xiong, Hui; Zheng, Yu
• Is Part Of: Biomedical signal processing and control. Volume 72(2022)Part B
• Publication Date: 2022-02
• Subjects: Mathematical morphology -- Adaptive weight filtering -- Transcranial magnetic stimulation signal -- Signal denoising; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2021.103348
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: Multi-channel transcranial magnetic stimulation signals denoising. New Application of Mathematical Morphology Algorithm. Signal matrix denoising reduces processing time. Multiple sets of structural elements to enhance filtering effect. Signal weighted combination to retain more useful information. Abstract: In order to solve the problem that with the interference of random noise and impulse noise to multi-channel transcranial magnetic stimulation (TMS) signals. In this paper a new morphological filtering algorithm is proposed based on improved generalized morphological filtering, which combines the basic operations of dilation, erosion, opening and closing operations, and constructs a new combined morphological filtering algorithm for multi-channel TMS signals denoising. Simulation results show that the proposed algorithm can not only remove impulse noise, but also suppress the white Gaussian noise effectively. Moreover, the proposed algorithm is more effective in reducing the processing time of multi-channel signals. Compared with generalized morphological filtering algorithm, the SNR of the proposed algorithm is improved by 23.93%–56.78%, the RMSE is reduced by 33.71%–50%, the MAE is reduced by 35.21%–38.89%, and the time consumed by the proposed algorithm is reduced by 88.24%. Finally, we use field data examples to demonstrate the successful performance of the proposed algorithm.