Chinese Journal of Biomedical Engineering

Home

Editorial Board

Download

Quick Search

Office Online

Current Issue

Download Articles

Download

Links

2022 Vol. 41, No. 4 Published: 2022-08-20

Reviews Communications Regular Papers

Regular Papers

	385	The Influence of tDCS on the Effective Brain Networks Based on Biased Partial DirectedCoherence
		Jin Ronghang, Luo Zhizeng, Shi Hongfei
		DOI: 10.3969/j.issn.0258-8021.2022.04.001
		Using brain network to study the effects of tDCS on brain function mechanisms and cerebral cortex state is of vital significance. In this study, we constructed effective brain networks of motor imagery in the different tDCS paradigms based on partial directional coherence. Taking the inflow and outflow rate of functional brain channel information of the effective brain networks as local features, and the average clustering coefficient and global efficiency as global features, we analyzed the influence of tDCS on the brain network characteristics of motor imagery. We found out that when the subjects performed left-hand motor imagery, the information outflow rate, inflow rate, average clustering coefficient and global efficiency of C4 channels after sham and tDCS anodal C4 stimulation was 0.142±0.014, 0.193±0.013, 0.585±0.046, 0.347±0.031 and 0.223±0.025, 0.258±0.023, 0.817±0.021 and 0.491±0.091, respectively, with significant differences (P<0.05). The information outflow rate, average clustering coefficient and global efficiency of C4 channel after tDCS cathodalC4 stimulation was 0.109±0.009, 0.356±0.037 and 0.252±0.024, respectively, which were significantly different from those of sham stimulation (P<0.05). The information inflow rate of C4 channel was 0.184±0.008, which was not significantly different from that of sham stimulation (P>0.05). The results indicated that the anodal tDCS effectively activated the activity of the cerebral cortex, made the brain area information exchange more frequently, increased the aggregation degree of the brain network, and improved the connectivity of the brain network. Cathodal tDCS would inhibit the activity of the cerebral cortex, reduce the outflow of brain information, reduce the aggregation degree of the brain network, and reduce the connectivity of the brain network.
		2022 Vol. 41 (4): 385-392 [Abstract] ( 441 ) HTML (1 KB) PDF (5736 KB) ( 335 )

	393	Seizure Detection in Focal Epileptic Patients Based on Adaptive Multi-Scale Brain Functional Connectivity
		Xu Jiayang, Yang Tingting, Li Wen, Li Kuo, Du Changwang, Liu Xiaofang, Sheng Duozheng, Yan Xiangguo, Wang Gang
		DOI: 10.3969/j.issn.0258-8021.2022.04.002
		Long-term EEG has been widely used to detect epileptic seizures in clinical practice. However, this approach is tedious and time-consuming, and largely depends on clinicians′ experience and subjective judgment. As a result, the accuracy and repeatability of the manual detection results are low. In this study, with the aim of solving the problem by using long-term EEG to monitor epileptic seizures, we proposed a novel adaptive and multiscale brain functional connectivity (AMBFC) method for epilepsy detection. Samples of 10 epilepsy patients during the seizure and non-seizure periods were selected as the research subjects. First, within a sliding time window, seven IMF components and residuals of the 19-channel EEG signal were extracted by MEMD. Then MVAR model was established to extract the outflow information intensity by the directional transfer function, and the features were combined and dimensionally reduced using PCA. Finally, the CSVM model was used to classify the seizure phase and non-seizure phase EEG, and the effect of epileptic seizures was evaluated through five-fold cross-validation. The results showed that the average accuracy rate of AMBFC algorithm for detecting EEG seizures was 98.6%, accuracy rate was 81.9%, recall rate was 81.4%, and F-measure value was 0.80. Compared with the detection results of the different IMF components, DTF-CSVM algorithm and methods in recent literatures, AMBFC algorithm was better. Except for the high accuracy, the proposed algorithm also achieved high precision, recall and F2 value. In conclusion, this method can be applied to real-time monitoring of long-term EEG.
		2022 Vol. 41 (4): 393-401 [Abstract] ( 309 ) HTML (1 KB) PDF (4333 KB) ( 315 )

	402	A Channel Selection Method Using EEG Signals for Fatigue Driving Detection
		Zheng Yun, Ma Yuliang, Sun Mingxu, Shen Tao, Zhang Jianhai, She Qingshan
		DOI: 10.3969/j.issn.0258-8021.2022.04.003
		Aiming at the problems of data redundancy and complex hardware facilities in the traditional fatigue detection method based on EEG, in this work, a channel selection method based on threshold filtering was proposed and used to select the optimal channel for several features by calculating the standard deviation and other metrics. Furthermore, hierarchical extreme learning machine (H-ELM) and PSO-H-ELM optimized by particle swarm optimization (PSO) were used to classify the selected optimal channel data and compared the accuracy to the results obtained by all the channels. The proposed method was validated using two groups of experimental data (one dataset was collected by a laboratory driving simulation device with 6 subjects and one dataset was a public dataset with 12 subjects and different collection devices). The results showed that for 18 subjects, many optimal channels were obtained from the ensemble empirical mode decomposition (EEMD) by using the power spectral density (PSD) feature of intrinsic mode function (IMF), and the distribution of optimal channels was roughly the same and the number of channels was small for the same device (8 and 11 channels, respectively). In addition, this channel selection method also significantly improved the accuracy of fatigue driving detection (the average accuracy of 18 subjects using the optimal channels was 99.75 %, 19.36 % higher than that using the full channel). Moreover, the ideal channel of sample entropy (SE) and the ideal channel of PSD hardly overlap, indicating that the two features were well complementary, and the practicality of proposed method is greatly improved by combining these 2 features. The results proved the effectiveness of the proposed method, which was of value in the application of fatigue driving detection.
		2022 Vol. 41 (4): 402-411 [Abstract] ( 454 ) HTML (1 KB) PDF (1506 KB) ( 510 )

	412	Screening of Wireless Capsule Endoscopy Image Sequence Based on Topic Model
		Nong Guixian, Pan Ning, Lu Heng, Hu Huaifei, Liu Haihua
		DOI: 10.3969/j.issn.0258-8021.2022.04.004
		Wireless capsule endoscopy (WCE), as a novel technology used to record images of the patient′s digestive tract, has greatly helped in the diagnosis of digestive tract diseases. However, during the detecting process, about 50,000-80,000 of images are produced for each patient, containing many disturbing images such as bubbles and impurities that greatly affect the efficiency of disease diagnosis. Most of the existing image screening methods, which are usually unstable and poorly generic, only have targeted on bubble images. Therefore, this paper proposed a topic model-based semantic analysis method to screen disturbing images from WCE sequences. The method firstly built an asymmetric autoencoder for image feature extraction, and used K-means algorithm to cluster the features of image patches in the training set to construct visual words; Secondly, the features of image patche in testing set were mapped into visual words to obtain the word frequency matrix of test images, resulting in the semantic representation of images based on visual words; Finally, the topic model was used to analyze the word frequency matrix and obtain the semantic classification of images. In this paper, the WCE dataset was obtained from 30 different patients and images in this dataset were annotated by a clinician with rich clinical experience. This dataset included 3 340 bubble images, 3 330 impurity images and 3 330 normal images, which were randomly divided into training and test sets in the ratio of 1∶1 for 10 times cross-validation. The experimental results showed that the proposed method effectively screened out the disturbing images, and the convolutional autoencoder based on the deep learning outperformed the traditional feature extraction method, obtaining 96.87% accuracy and effectively improving the efficiency of disease diagnosis.
		2022 Vol. 41 (4): 412-419 [Abstract] ( 171 ) HTML (1 KB) PDF (7176 KB) ( 47 )

	420	Finger Vein Recognition Based on Curve Descriptor
		Su Dan, Wang Xinqiang, Liu Yuhang, Lu Yaopeng, Li Ting, Nie Zedong
		DOI: 10.3969/j.issn.0258-8021.2022.04.005
		Finger vein recognition has become a research hotspot in the field of identity recognition due to its advantages of high anti-counterfeiting, uniqueness, stability, and liveness detection. At present, most recognition algorithms based on vein structure features consider the feature of detail points, but it is easy to ignore the curve feature of the vein network structure, which would cause the loss of part of structural information and affect the recognition results. To solve the problems, this paper proposed a finger vein recognition algorithm based on curve descriptor. Firstly, the skeleton structure of finger veins was extracted, the intersection points and endpoints of the veins were detected, and the intersection points and endpoints were used to divide the vein skeleton into several curve segments. Secondly, the curve arc descriptors and intersecting arc descriptor were proposed based on the relative position and shape characteristics of the intersection points and curve segments, and the structural feature matrix of finger veins was extracted. Finally, the matching intersecting arc pair was calculated according to the weighted distance formula proposed, and then the degree of image matching was judged. Experiments were carried out on a finger vein database with a sample size of 840 images. The experimental results showed that the equal error rate of the LBP, LTP and SURF algorithms was 4.47%, 3.99% and 6.08% respectively, while the equal error rate of the proposed method was only 1.63% that was lower than that of LBP, LTP and SURF algorithms, indicating that the method in this paper had a certain universality and application prospect in finger vein recognition.
		2022 Vol. 41 (4): 420-430 [Abstract] ( 193 ) HTML (1 KB) PDF (3170 KB) ( 264 )

	431	Multi-Category Intestinal Polyp Image Classification Network Based on Edge Prior Information
		Li Sheng, Cao Jing, Ye Shufang, Dai Fei, He Xiongxiong
		DOI: 10.3969/j.issn.0258-8021.2022.04.006
		The classification of intestinal polyps can help endoscopists to assist in diagnosis and distinguish between high-risk polyps requiring immediate treatment and low-risk polyps that can be deferred. Existing polyp classification algorithms based on deep learning can′t distinguish the high degree of inter-class similarities images, and need to be improved for multi-category polyp classification task. In this paper, a multi-category polyp image classification network based on edge prior information was proposed, including edge detection stage, edge feature descriptor extraction stage and polyp classification stage. Firstly, at the skip connection layer in the edge detection stage, a reverse attention edge monitoring module was designed and embedded to better capture the details of polyp edge. Secondly, under the guidance of the prior knowledge of the endoscopist, the perimeter size was represented by counting the number of pixels on the edge of the polyp, and the concavity and convexity were used to represent the smoothness feature, so as to supplement the insufficiency of neural network feature extraction. Finally, the channel attention was inserted after DenseBlock4 of the classification network to adaptively capture discriminative features. The private dataset was consisted of 1 050 desensitized original images that are collected from the Digestive Endoscopy Center of Lishui People′s Hospital within the year 2018 to 2019. Five-fold cross-validation was conducted in the polyp four-category dataset constructed in this paper, and the overall accuracy reached 77.29%, which was 6.46% higher than the best results of existing algorithms. The classification network fused with edge prior information can effectively discriminate two groups of polyp images with high degree of inter-class similarities, namely non-adenomatous polyps and low-grade adenomatous polyps, high-grade adenomatous polyps and adenocarcinoma. The established network in this paper increased the robustness and improved classification performance, providing auxiliary opinions for doctor diagnosis under limited training dataset.
		2022 Vol. 41 (4): 431-442 [Abstract] ( 247 ) HTML (1 KB) PDF (7387 KB) ( 191 )

	443	Object Detection of Pneumonia Images Based on Deep Learning
		He Di, Liu Lixin, Liu Yujie, Xiong Feng, Qi Meijie, Zhang Zhoufeng
		DOI: 10.3969/j.issn.0258-8021.2022.04.007
		Pneumonia is a disease that seriously endangers people′s health. Lung X-rays are usually used for pneumonia examination. The diagnosis of pneumonia is a very important step before the treatment of pneumonia. However, due to the interference of other lung diseases, the explosion of medical data, and the lack of professional pathologists, it is very difficult to accurately diagnose pneumonia. Deep learning can imitate the mechanism of the human brain to interpret medical image datasets with improved accuracy and efficiency, therefore, has been widely used in pneumonia image detection. In this paper, three deep learning-based object detection models, SSD, faster-RCNN and faster-RCNN optimization model, were used to study 26 684 labeled lung X-ray images from the Kaggle dataset. The original X-ray images were preprocessed and then input into the three deep learning models to detect single or two lesion areas. The performance of the three models was evaluated and compared through loss function, classification accuracy, regression accuracy and number of mis-detected lesions by testing 500 randomly selected images. The results showed that faster-RCNN performed better than SSD in performance metrics; Faster-RCNN optimization model was superior to the other two models with the loss value that was small and could quickly reach stability, the average classification accuracy of 93.7%, the average regression accuracy of 79.8% and the number of mis-detected lesions of 0, which would be helpful for the accurate identification and diagnosis of pneumonia.
		2022 Vol. 41 (4): 443-451 [Abstract] ( 341 ) HTML (1 KB) PDF (7432 KB) ( 217 )

	452	A Rapid Electrochemical Detection Method of Low-Concentration Dopamine Based on Machine Learning
		Liu Zhe, Sun Lesheng, Yu Jun, Lu Ning, Xu Ying, Guo Miao
		DOI: 10.3969/j.issn.0258-8021.2022.04.008
		Portable potentiostat is facing the problem that the accuracy is low and vulnerable under the test conditions. This paper proposed an electrochemical detection data analysis method by combining nanomaterial modified electrodes and machine learning. Under the presence of multiple experimental interference factors, this method was able to achieve the accurate detection of dopamine (DA). The AuNPs/GCE electrode was prepared by electrodepositing gold nanoparticles on the surface of glassy carbon electrode (GCE) by chronoamperometry, and the electrocatalytic activity for the redox of dopamine of AuNPs/GCE was verified by cyclic voltammetry (CV). Under the different solution pH values and scanning speeds, the AuNPs/GCE electrode was applied to perform repetitive accurate cyclic voltammetric detection of dopamine solutions of different concentrations. After the extraction of important characteristics including the peak height, peak potential, baseline slope, peak area and initial redox potential of the detection data, the extreme gradient boosting tree model (XGBoost) and the random forest model (RF) were applied to construct a two-stage concentration prediction analysis. The results showed that the MAE, RMSE and MAPE% of XGBoost-RF concentration prediction model were reduced to 53.9%, 39.7% and 2.7% respectively compared with the traditional SVR model. The training time of RF prediction model was reduced by 23%, the prediction accuracy was improved by 7%, and the fitting degree (R-squared) between predicted value and experimental value was 0.943. In conclusion, this method effectively reduced the influence of different experimental factors in the detection process. It Also improved the detection accuracy and reduced the complexity of the experiment. Therefore, it is of great significance to realize the on-site and rapid electrochemical detection of microscale element.
		2022 Vol. 41 (4): 452-461 [Abstract] ( 228 ) HTML (1 KB) PDF (11023 KB) ( 163 )

Reviews

	462	Applications of Spectrum Analysis Technology Based on Time-Domain Photoacoustic Signalin Biomedical Field
		Zheng Jiaxin, Tian Rui, Liu Mingqing, Zan Kehua, Wang Yihan, Zhu Shouping
		DOI: 10.3969/j.issn.0258-8021.2022.04.009
		Time-domain photoacoustic (PA) signal measurement and spectral analysis technique is a non-invasive detection method that can provide the structural and functional information of biological tissue. Combining with the high contrast of optical modality and the high resolution of ultrasonic modality in deep tissue, PA signal data sets of target tissues under different wavelengths of light excitation are processed and analyzed. Compared with the conventional spectral detection, this technique is less susceptible to the limitations of the shape and morphology of the object to be measured and is not affected by the light scattering, therefore has a high sensitivity for the detection of deep tissues. In contrast to photoacoustic imaging, this technique does not require image reconstruction and focuses on achieving quantitative analysis. This article summarized the applications of time-domain PA signal spectral analysis technique in the detection of biological tissue, biological fluid, and biological exhaled gas, and reviewed the research progress and development directions of this technique around the improved experimental systems or different signal processing methods used in various studies.
		2022 Vol. 41 (4): 462-472 [Abstract] ( 218 ) HTML (1 KB) PDF (4061 KB) ( 370 )

	473	Current Status and Prospects of Diagnostic Methods for Diabetic Lower-Extremity ArterialDisease
		Yang Yuxiang, Shi Zhuangzhi, Zhang Fu, Zhang Linshan, Zhu Yan, Li Zhonglin
		DOI: 10.3969/j.issn.0258-8021.2022.04.010
		Diabetes mellitus (DM) lower extremity arterial disease (LEAD) is one of the main pathological factors leading to diabetic foot (DF) that is the most serious chronic complications of DM with the highest treatment cost. Therefore, timely diagnosis of LEAD is crucial to prevent the occurrence of DF. This paper summarized the LEAD diagnostic methods commonly used in clinic, including subjective evaluation, image detection and physiological parameter detection, and expounded the basic principles, advantages and disadvantages of these methods. In view of the current increasing demand for early diagnosis of LEAD, the research progress of microvascular disease (MVD) and local arterial pulse wave velocity (PWV) detection technology and their feasibility of clinical applications were summarized, which is expected to become a breakthrough for early diagnosis of asymptomatic LEAD in DM patients.
		2022 Vol. 41 (4): 473-484 [Abstract] ( 234 ) HTML (1 KB) PDF (969 KB) ( 250 )

	485	Research Progress for the Analysis of Images and Genetic Features in Alzheimer′s Disease
		Han Liting, Yao Xufeng, Jin Yu, Zhao Congyi, Huang Gang
		DOI: 10.3969/j.issn.0258-8021.2022.04.011
		Alzheimer′s disease (AD) is one of the most common neurodegenerative diseases, and its phenotype has shown susceptible to genetic factors. In recent years, with the wide application of multimodal brain imaging and high-throughput genomics in medical imaging, it has become a new hotspot to explore the association analysis between images and genes by means of data mining and mathematical modeling. Till now, the combined analysis of images and genetic characteristics has been used to study AD and has made significant progress in the early diagnosis, classification, and prognostic analysis. This article first summarized the imaging and genetic features, then explained the application of statistics and machine learning (ML) methods in the joint analysis of image gene features, and finally summarized and proposed its development perspectives.
		2022 Vol. 41 (4): 485-492 [Abstract] ( 199 ) HTML (1 KB) PDF (828 KB) ( 387 )

	493	Biomechanical Factors in Space Flight-Associated Neuro-Ocular Syndrome
		Wang Xiaofei, Liu Tingting, Fan Yubo
		DOI: 10.3969/j.issn.0258-8021.2022.04.012
		Long duration space flight can induce structural changes of eye and corresponding vision loss, which is defined as space flight-associated neuro-ocular syndrome (SANS). SANS is the main problem of eye for future human space exploration. Up to date, mechanisms of SANS are still unknown, which impedes the development of countermeasures and the selection criteria of astronauts. SANS is correlated with optic nerve subarachnoid space cerebrospinal fluid pressure elevation in microgravity environments. However, knowledge on the detailed response of eyes to elevated pressure and its corresponding countermeasures is lack. This paper summarized the morphological changes of the optic disc, posterior segment of the eye, subarachnoid space and optic nerve caused by SANS. Next, the pathogenesis of SANS proposed in the literature, the factors affecting the susceptibility of SANS and the countermeasures were reviewed and the relationship between SANS and the optic nerve subarachnoid space cerebrospinal fluid pressure in microgravity environment was summarized. At last, we discussed how to effectively combine ground-based and space-based experiments to explore the pathogenesis of SANS.
		2022 Vol. 41 (4): 493-501 [Abstract] ( 253 ) HTML (1 KB) PDF (4645 KB) ( 338 )

	502	Application of Balloon in Angioplasty for Calcified Coronary Artery
		Li Jiasong, Lin Changyan
		DOI: 10.3969/j.issn.0258-8021.2022.04.013
		Severe calcified coronary artery lesions are one of the most challenging lesions during percutaneous coronary intervention. Calcification often impedes balloon dilation and has the stent beams and polymer coatings damaged, leading to poor clinical outcomes or even failure to complete the surgery. Therefore, preprocessing of calcified plaques is necessary, and how to select the right coronary angioplasty balloon for the preprocessing of different calcified lesions has become a new research topic. Based on structure information, application status, and computer simulation experiments, this paper summarized the balloon applied in angioplasty for calcified lesions in percutaneous coronary intervention including cutting balloon,lacrosse NSE balloon. The value of the computer simulation experiment in guiding the balloon selection during percutaneous coronary endovascular angioplasty for calcified plaques was proposed as well.
		2022 Vol. 41 (4): 502-507 [Abstract] ( 208 ) HTML (1 KB) PDF (778 KB) ( 434 )

Communications

	508	Face Video Heart Rate Detection Based on FastICA and ICEEMDAN
		Zhao Mingkang, Wang Zhen, Qi Chencheng, Wang Yixiao, Zhang Shuai
		DOI: 10.3969/j.issn.0258-8021.2022.04.014

		2022 Vol. 41 (4): 508-512 [Abstract] ( 207 ) HTML (1 KB) PDF (1158 KB) ( 237 )

Copyright © Editorial Board of Chinese Journal of Biomedical Engineering
Supported by:Beijing Magtech