Chinese Journal of Biomedical Engineering

Home

Editorial Board

Download

Quick Search

Office Online

Current Issue

Download Articles

Download

Links

2018 Vol. 37, No. 4 Published: 2018-08-20

Reviews Communications Regular Papers

Regular Papers

	385	A New Fully Convolutional Network for 3D Liver Region Segmentation on CT Images
		Sun Mingjian, Xu Jun, Ma Wei, Zhang Yudong
		DOI: 10.3969/j.issn.0258-8021.2018.04.001
		Liver segmentation has important clinical value in liver tumor resection and liver transplantation volume measurement. Because the intensity value of liver and adjacent organs is very close in CT images, the three-dimensional (3D) automated segmentation of the liver region is a challenged task. In order to make the accurate segmentation of liver region, a new deep fully convolutional network (FCN) structure 3DUnet-C2 was proposed. This network made full use of the three-dimensional spatial information of CT image, and combined well the characteristics of shallow and deep layers. In particular, a new network training strategy was proposed. The primary model was obtained by selecting the clear image and intercepting the liver region as a sample. Then the model was leveraged to initialize the network parameters so that the network can converge. Finally, on the basis of the original model, the 3DUnet-C2-CRF model was constructed by using the three-dimensional conditional random field to optimize the liver segmentation boundary. In order to verify the performance of the proposed 3DUnet-C2-CRF on 3D segmentation of liver regions, 100 CT images were chosen from the data set of the ISBI2017 Liver Tumor Segmentation Challenge. The Dice coefficient of the segmentation accuracy of the 3DUnet-C2-CRF model on 20 test images reached 96.9%, which is higher than the Dice coefficient of 3DUnet and Vnet models. Experimental results showed that the 3DUnet-C2-CRF model had better feature expression capability and more generalization performance, which improved the segmentation accuracy of the model.
		2018 Vol. 37 (4): 385-393 [Abstract] ( 629 ) HTML (1 KB) PDF (14147 KB) ( 225 )

	394	DTI Image Segmentation Algorithm Based on the Improved Spatial Fuzzy Clustering
		Liu Xuyu, Zhang Xiangfen, Ma Yan, Li Chuanjiang, Yang Yanqin
		DOI: 10.3969/j.issn.0258-8021.2018.04.002
		Aiming to resolve the problems of initial clustering selection randomness and noise sensitivity of fuzzy C means algorithm, this paper proposed an image segmentation algorithm based on the improved spatial fuzzy clustering to segment the DTI image of human brain. In this paper, we used the local density kernel function and the center distance function to select the initial clustering center accurately, which not only solved the problem of clustering effect instability caused by random selection of cluster center, but also made the objective function converge quickly, and improved the segmentation efficiency. Moreover, the proposed algorithm reduced the influence on the segmentation result caused by image noise and human factors by integrating normal distribution spatial information into fuzzy membership function. We segmented DTI data of human brain with the proposed method, FCM and SFCM to evaluate the clustering effect of the algorithm. In the experiments, following data were employed, including segmented 58 cases of DTI data provided by the University of Minnesota Biomedical Functional Imaging and Nerve Engineering Laboratory, 3 cases of FA parameter images, and 6 cases of iterative noisy human brain DTI images. Results show that the segmentation coefficient of proposed algorithm reached 0.9841. In the same image, the algorithm obtained the most improvement of 20.2% than FCM on the partition coefficient, and the most decline of 19.8% than SFCM on the partition entropy; The average number of iterations of the algorithm was 32, which is significantly lower than 52 of FCM and 76 of SFCM. Therefore, the algorithm can segment the important target accurately and quickly, and the segmentation results are insensitive to image noise.
		2018 Vol. 37 (4): 394-403 [Abstract] ( 416 ) HTML (1 KB) PDF (19822 KB) ( 146 )

	404	Diagnosis of Breast Cancer Based on Tumor Parameters and Visualization of the Attribute Partial Order Structure Diagram
		Liang Huaixin, Song Jialin, Zheng Cunfang, Hong Wenxue
		DOI: 10.3969/j.issn.0258-8021.2018.04.003
		In order to realize the visualization of the rules of breast cancer data, a method based on the combination of Lasso and incremental learning, was proposed, using the optimized attribute partial order structure diagram as a tool. Firstly, having the dimensions reduced by using Lasso to select the features of the breast cancer data, and four attributes that gained the largest correlation were selected from nine features. Granulation process was completed under the Gini index, generating the formal context by means of the incremental learning algorithm. Next, the second Lasso process was completed, which made the dimensions reduced from 17 to 3. Meanwhile, a new method processing the rows and columns of the formal context based on the Gini index and the covering theory was proposed to generate the attribute partial order structure diagram to visualize the rules concerned. As there have been seven rules extracted by analyzing the diagram reported in literatures,we compared the proposed classification accuracy of the method with those classical mainstream classifiers. Results showed that the classification precision of our method reached 96.52%, higher than the other five classifiers including Random Forest (94.25%), Adaboost (90.00%), 1NN (91.33%), 3NN (90.67%), and SVM (95.00%). At last, different incremental proportional (10%-90%) data were used to verify the effect of incremental learning algorithm, results showed that the model had been completed when the amount of data reached 30%, and the precision was almost approaching to that of support vector machine, which proved that the proposed method represented an effective means of visualizing the diagnosis rules of breast cancer.
		2018 Vol. 37 (4): 404-413 [Abstract] ( 523 ) HTML (1 KB) PDF (969 KB) ( 437 )

	414	Classification of Breast Mass in 3D Ultrasound Images with Annotations Based on Convolutional Neural Networks
		Kong Xiaohan, Tan Tao, Bao Lingyun, Wang Guangzhi
		DOI: 10.3969/j.issn.0258-8021.2018.04.004
		The automatic classification of breast tumor in ultrasound images is of great significance to improve doctors' efficiency and reduce the rate of misdiagnosis. The novel 3D breast ultrasound data contains more information for diagnosis, but images from different directions have their distinct performance as a result of this ultrasound imaging mechanism. For this breast ultrasound data, this paper designed three kinds of convolutional neural network model using its flexibility and characteristic of learning automatically, and the three models were able to accept transverse plane images, transverse plane and coronal plane images, images and annotations information. The effects of different information fusion on the accuracy of breast tumor classification were investigated. A dataset contains 880 images (i.e., 401 benign images, 479 malign images) and their annotations were employed, and we performed 5-fold cross validation to calculate the accuracy and AUC of each model. The experimental results indicated that the models designed in this paper can deal with the images and annotations simultaneously. Compared with the single-input model, the multi-information fusion model improved the accuracy of classification by 2.91%, and achieved the accuracy of 75.11% and AUC of 0.829 4. The proposed models provided a reference for the classification application of convolutional neural networks with multi-information fusion.
		2018 Vol. 37 (4): 414-422 [Abstract] ( 770 ) HTML (1 KB) PDF (4269 KB) ( 847 )

	423	Multi-Label Recognition of Pedestrian Attributes Based on Deep Learning
		Li Yapeng, Wan Suiren
		DOI: 10.3969/j.issn.0258-8021.2018.04.005
		Pedestrian attributes usually refer to some of the external characteristics of pedestrians that can be observed, such as gender, age, clothing type, carrying objects, etc. As soft biological features of pedestrians, pedestrian attributes are very important for pedestrian detection and re-identification, and show great potential in intelligent video surveillance scenarios and video based business intelligence applications. Among the current multi-label classification methods of pedestrian attributes, two of them are mainly employed, one is based on handcrafted features and the other is based on the deep learning methods. However, the methods of handcrafted features are difficult to deal with complex real video surveillance scenes, results obtained in practical applications are not ideal. In this paper we used a deep convolutional network model with three convolutional layers and two full-connected layers. Using the Sigmoid cross-entropy loss function, the training platform was the Caffe deep learning framework, the dataset used was PETA containing 19,000 pedestrian images. Ten kinds of pedestrian attributes were trained and tested, and an average recognition accuracy of 85.2% was reached. After adding the positive sample proportional exponential factor to improve the loss function, the average recognition accuracy reached 89.2%, which significantly improved the performance of the network.
		2018 Vol. 37 (4): 423-428 [Abstract] ( 1160 ) HTML (1 KB) PDF (3327 KB) ( 633 )

	429	Analysis of Parameter Perfusion Imaging for Contrast-Enhanced Ultrasound Based on S-G Filter
		Wang Bengang, Ding Hong, Peng Shiyun, Xu Zhiting, Fu Tiantian, Wang Wenping
		DOI: 10.3969/j.issn.0258-8021.2018.04.006
		The previous perfusion analysis software for contrast-enhanced ultrasound (CEUS) provided quantitative parameters under the assumption that the region of interest (ROI) was a whole homogeneous perfusion area. To overcome the imperfection, we herein presented a fast data smoothing method with time domain based on S-G filter of ROI’s internal regional perfusion imaging based on pixel parameter analysis. The time intensity curve (TIC) of each pixel was obtained from the original DICOM dynamic imaging of CEUS data, the quantitative analysis of parameters of TIC were conducted based on the S-G filter, including time of arrival (AT), time to peak (TTP), peak intensity (PI), mean transit time (MTT), the area under the curve (AUC), and rising slope (RS). After that the pixel perfusion parameters coding was displayed in the color mode accurately and intuitively. The stability test was performed on 15 cases of hepatocellular carcinomas to obtain different quantitative parameters of TIC by three investigators respectively. The intraclass correlation coefficient of all parameters was more than 0.80.
		2018 Vol. 37 (4): 429-434 [Abstract] ( 494 ) HTML (1 KB) PDF (16493 KB) ( 170 )

	435	Investigation on Brain-Heart Coupling during Head-Up Tilt Test
		Dang Shijie, Wang Xiaoni, Liu Binbin, Zhang Jianbao
		DOI: 10.3969/j.issn.0258-8021.2018.04.007
		Cerebral cortex can modulate cardiovascular activities, which is mediated by autonomic nervous system (ANS). The investigation of the underlying mechanism plays an important role in the diagnosis, treatment, prevention and control of diseases such as arrhythmia, hypertension and heart failure. ECG, BP, chest impedance signals and near infrared spectroscopy (NIRS) signals in the medial prefrontal cortex (MPFC) were collected synchronously in 27 healthy subjects in supine condition and during 70° head-up tilt test (HUT). Heart rate variability (HRV) and Granger causality (GC) methods were applied to investigate cardiovascular responses in HUT and demonstrate the interactions between cerebral blood volume (CBV) in the MPFC and sympathetic/parasympathetic outflows (indicated by nLF and HF, respectively) as well as between CBV in the MPFC and RRI. Results showed that 1) CO decreased by 24% (P<0.001), HR and DBP increased by 35% (P<0.001) and 7% (P<0.001) respectively, while SBP almost showed no change compared with baseline even though it declined at the beginning of HUT; 2) nLF increased while HF decreased gradually, and then kept a relatively stable state with the former more activated (38.75±9.25 vs 58.62±8.58,P<0.001) while the latter less activated (1.11±0.76 vs 0.80±0.48,P<0.05) than the supine condition; 3) CBV in the MPFC was significantly correlated with both sympathetic and parasympathetic outflows and the correlation coefficients were 0.58 and 0.66 respectively; 4) GC showed frequency specificity with preference of “brain→heart” (P<0.05) in the frequency band of 0.04-0.15 Hz while “heart→brain” (P<0.001) in the frequency band of 0.15-0.4 Hz. Our results indicated that significant cardiovascular responses, mainly modulated by sympathetic and parasympathetic outflows, were elicited during HUT and that the MPFC was involved in the ANS-mediated cardiovascular control with ～0.1 Hz oscillations mediating the process. This study provided new evidence for the brain-heart coupling (the characteristic brain-heart interactions in healthy subjects during HUT) and we found that ～0.1 Hz oscillations play an important role in mediating cerebral autonomic control.
		2018 Vol. 37 (4): 435-444 [Abstract] ( 432 ) HTML (1 KB) PDF (1116 KB) ( 431 )

	445	Research on Delta-Gamma Phase Amplitude Coupling Based on Mental Fatigue
		Yang Shuo, Ji Yakun, Wang Lei, Hao Pengru, Xu Guizhi
		DOI: 10.3969/j.issn.0258-8021.2018.04.008
		Mental fatigue is a state that a person cannot concentrate and finish the task efficiently because of long-term highly concentration on the work, which has extremely influenced upon the health and life of people. However, there have been no simple and effective detection methods.Meanwhile recent studies have found out that phase amplitude coupling (PAC) between low and high rhythm may be related to the information integration in cognitive activities, providing new messages for the detection of mental fatigue.In this paper, the phase amplitude coupling between delta and gamma rhythm was used to study the EEG data recorded before and after mental fatigue. Neuroscan EEG system was used to collect data, and phase amplitude coupling of 14 subjects were calculated and analyzed by paired t test. The results showed that the delta phase of over 90% electrodeson the whole brain area jointly modulated gamma amplitude of frontal lobe, occipital lobe, parietal lobe and frontal lobe, and the coupling effect among different brain regions significantly increased in the first three subareas while the last region significantly decreased when participants were mental fatigued. This study showed that phase amplitude coupling could predict the change of macroscopic behavior caused by mental fatigue and provide a new index for the detection of mental fatigue.
		2018 Vol. 37 (4): 445-450 [Abstract] ( 593 ) HTML (1 KB) PDF (3682 KB) ( 572 )

	451	Extraction of Entity Interactions Based on Multiple Feature Fusion Linear Kernel SVM Approach
		Wei Xing, Hu Dehua, Yi Minhan, Chang Xuelian, Yang Xiaodi, Zhu Wenjie
		DOI: 10.3969/j.issn.0258-8021.2018.04.009
		Improving the performance of interaction mining algorithm can help to explore some innovative ideas in the biomedical literature. We proposed a novel feature-based linear kernel support vector machine (SVM) approach to extract and investigate the interactions between diabetes mellitus, genes and drugs. We elaborated the five types of features (entity, entity pair, dependency graph, parse tree, noun phrase-constrained coordination) used, including two novel features, word pair and noun phrase-constrained coordination features. Then 173 interactions between 13 kinds of diabetes mellitus and 23 genes, 79 interactions between 13 kinds of diabetes mellitus and 26 drugs, 159 interactions between 18 genes and 17 genes, 619 interactions between 8 kinds of diabetes mellitus, 23 genes and 26 drugs were ontained. And 27 new entity interactions were predicted. After that we constructed the interaction network of the disease-gene, gene-drug, and disease-gene-drug. The experimental results showed that the proposed method was comparable with the algorithms used in CoPub (0.710), PubGene (0.609), FBK-irst (0.547, 0.800) and WBI (0.510, 0.759), the highest accuracy increased by about 5% (0.847 vs 0.800, and the minimum increased by over 20% (0.742 vs 0.510), which provided perspectives for applications of biomedical big data.
		2018 Vol. 37 (4): 451-460 [Abstract] ( 358 ) HTML (1 KB) PDF (8874 KB) ( 215 )

	461	Three-Dimensional Reconstruction and Finite Element Analysis of Rat Trabecular Meshwork Basing on Two-Photon Microscope Images
		Zhang Jing, Qian Xiuqing, Zhang Haixia, Ren Lin, Liu Zhicheng
		DOI: 10.3969/j.issn.0258-8021.2018.04.010
		Glaucoma is a disease of irreversible blindness with the most risk factors of ocular hypertension that causes damage to the visual function and optic nerve. So far, clinical evidence indicates that increased intraocular pressure (IOP) is a function of elevated resistance to drainage of aqueous humor, especially the increase of outflow resistance within the trabecular meshwork (TM). Therefore, it is very important to figure out what effects of high IOP on the structures of TM and analyze the morphologic changes of TM contributing to the increase of outflow resistance in turn. In this study, animal model of acute IOP of rat was established by methods of anterior chamber perfusion. Six SD rats were divided into A, B two groups. After being killed, the left eyes of the SD rats in group B were perfused at pressure of 60 mmHg for imaging under the condition of ocular hypertension, and the other eyes was used as the control group. The tomographic sequence images of TM were obtained in the condition of normal or high IOP by using two-photon microscopy. The effects of IOP on the TM porosity based on image processing were quantitatively investigated. The three-dimensional model of TM was obtained in the normal IOP eye, and the morphological changes of TM were analyzed under different intraocular pressures by using the finite element method. The effect of high IOP on the outflow resistance of TM was analyzed by using the experiment integrated with the simulation method. In the high IOP group, the TM beams became collapsed and merged with surrounding tissues, and the damage of TM tissue near the anterior chamber is more serious. Through the analysis of finite element method, it was found out that the biggest deformation was located in the region that the porosity of TM was larger. Meanwhile, it was indicated that the greater the pressure was, the greater the damages of TM beams was. Eyes with the high IOP may cause the structural changes of TM. Fibers of TM were collapsed. The possibility of increasing outflow resistance could occur in high IOP eyes but was much smaller in normal eyes.
		2018 Vol. 37 (4): 461-467 [Abstract] ( 408 ) HTML (1 KB) PDF (7056 KB) ( 340 )

	468	Precise Optimization of 3D Bio-Printing Cell-Laden Hydrogel Construct
		Huang Mengjie, Luo Li, Wang Ling, Xu Mingen
		DOI: 10.3969/j.issn.0258-8021.2018.04.011
		The combination of 3D bio-printing technology and hydrogel provides an attractive solution for manufacturing tissues and organs with complex structures and functions. The internal constructs can be custom-printed to imitate the 3D microenvironment of tissues and organs, which is critical for cell growth, tissue formation and regeneration after printed. However, precise matching between the printed and the design is still challenging due to variable physical and chemical properties of hydrogel. In this paper, an optimized method based on optical coherence tomography (OCT) for 3D bio-printing cell-laden hydrogel construct was proposed. The tissue was imaged by 3D non-destructive homemade OCT system, and quantitatively characterized, then the mismatch of structural parameters between the design and the printed was reduced using the experimentally obtained equations, and the accuracy and stability of 3D bio-printing were improved. The results demonstrated that the deviation of the key structural parameters between the printed and the design was controlled to less than 7%, which was much lower than the deviation about 40% reached by conventional 3D bio-printing. The controlled parameters included pore size, strut size, porosity, surface area and pore volume. And the cell viability of two weeks incubation was increased from about 80% to over 90%. It was concluded that online quantitative evaluation and feedback iterative analysis system based on OCT provided a potential tool for mass customization of cell-laden hydrogel tissues, 3D bio-printing tissues and organs.
		2018 Vol. 37 (4): 468-480 [Abstract] ( 399 ) HTML (1 KB) PDF (31791 KB) ( 204 )

	481	The Preparation of uPAR-Targeted MRI Probe and its Targetability to Breast Cancer Cells
		Yang Yang, Meng Jie, Wen Tao, Chen Bo, Liu Fei, Gu Ning, Xu Haiyan, Yu Wei, Liu Jian
		DOI: 10.3969/j.issn.0258-8021.2018.04.012
		Breast cancer is one of the most common malignancies for women. Early detection of primary and metastatic lesions of breast cancer is a key point to improve the survival rate of patients. Studies have shown that urokinase-type plasminogen activator receptor (uPAR) is highly expressed in breast cancer cells and tumor-associated cells, while lowly expressed in normal tissue cells. In this study, by conjugating amino-nitrilotriacetic acid (NTA) to the surface of superparamagnetic iron oxide nanoparticles (SPIO), we obtained an universal magnetic resonance imaging (MRI) module (SPIO-NTA). Then, amino-terminal fragment (ATF) of urokinase-type plasminogen activator (uPA), the native ligand of uPAR, was introduced to SPIO-NTA via histidine tags (His tags) to obtain an uPAR-targeted MRI agent (SPIO-ATF). The diameter of the γ-Fe₂O₃ core was less than 10 nm. The hydrodynamic diameter of SPIO-ATF was about 77 nm and Zeta potential was about -13 mV. The results of protein gel electrophoresis and Coomassie brilliant blue staining indicated the successful conjugation of ATF to SPIO. In vitro experiments showed a positive correlation between the amount of SPIO-ATF bound with 4T1 cells and the expression level of uPAR, indicating that SPIO-ATF can specifically bind to 4T1 cells via the interaction between ATF and uPAR. In addition, the cellular uptake of SPIO-ATF was increased with the increase of incubating concentration of SPIO-ATF. Under experimental conditions, neither SPIO nor SPIO-ATF showed obvious cytotoxicity. In conclusion, we developed a novel MRI probe for the target detection of breast cancer, providing a new strategy for the early diagnosis of breast cancer.
		2018 Vol. 37 (4): 481-488 [Abstract] ( 333 ) HTML (1 KB) PDF (3961 KB) ( 234 )

Reviews

	489	Review of Human Enhancement through Electrical Nerve Stimulation
		Pan Ziqi, Zhu Bingquan, Liu Jing
		DOI: 10.3969/j.issn.0258-8021.2018.04.013
		Recently there has been growing research on electrical nerve stimulation for human enhancement, including augmentations on cognition (memory, learning and emotion), perception and movement. In addition, brain machine interfaces (BMIs) stimulate human brain precisely. And a tremendous effort has been applied to boost BMI’s information gathering and processing capacity. Transcranial direct current stimulation (tDCS), representing non-invasive stimulation, implements easily but needs further investigation of long-term effectiveness and spatial resolution. Invasive stimulation methods, including deep brain stimulation (DBS) and vagus nerve stimulation (VNS), have been widely accepted in clinical practice with high spatial resolution. But these methods have the restriction for enhancement in both research and application. This article reviews the research progresses in electrical stimulation for human enhancement and advances in BMI, and discusses the limitations and tendency.
		2018 Vol. 37 (4): 489-497 [Abstract] ( 473 ) HTML (1 KB) PDF (884 KB) ( 1078 )

	498	Research Progress of Effects of Transcranial Magnetic Stimulation on the Depression Animal Model
		Wang Ling, Yang Jiajia, Wang Faqi, Wan Baikun, Ming Dong
		DOI: 10.3969/j.issn.0258-8021.2018.04.014
		Depression is a common mental illness that has aroused general concerns. The depression has the characteristics of high prevalence, high recurrence rate and high mortality. It’s important and of significanceto study the pathogenesis and effective treatment approches. As a non-invasive, safe and effective method, transcranial magnetic stimulation (TMS) has superiority in treating the depression. Although many clinical studies have proven its effectiveness,basic experimental studies on the animal model are still needed for fully insight into the optimal selection of stimulus parameters and its mechanism. Here we reviewed the research of the effects of TMS on depression rat model in recent years from two aspects: stimulus parameters and the underlying mechanisms. This paper discussed some major problems and future prospects of the research on TMS treating depression rat model in order to promote the development and application of TMS.
		2018 Vol. 37 (4): 498-507 [Abstract] ( 487 ) HTML (1 KB) PDF (1521 KB) ( 560 )

Communications

	508	The Development of a Dual Wavelength Optical Coupling Functional Imaging System for Breast Diagnosis
		Xu Ruiwen, Rong Meng, Li Kaiyang
		DOI: 10.3969/j.issn.0258-8021.2018.04.015

		2018 Vol. 37 (4): 508-512 [Abstract] ( 364 ) HTML (1 KB) PDF (1940 KB) ( 499 )

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