Publications by authors named "Jinzhu Yang"

29 Publications

  • Page 1 of 1

A review on segmentation of lung parenchyma based on deep learning methods.

J Xray Sci Technol 2021 Aug 28. Epub 2021 Aug 28.

Key Laboratory of Intelligent Computing in Medical Image, Ministry of Education, Shenyang, China.

Precise segmentation of lung parenchyma is essential for effective analysis of the lung. Due to the obvious contrast and large regional area compared to other tissues in the chest, lung tissue is less difficult to segment. Special attention to details of lung segmentation is also needed. To improve the quality and speed of segmentation of lung parenchyma based on computed tomography (CT) or computed tomography angiography (CTA) images, the 4th International Symposium on Image Computing and Digital Medicine (ISICDM 2020) provides interesting and valuable research ideas and approaches. For the work of lung parenchyma segmentation, 9 of the 12 participating teams used the U-Net network or its modified forms, and others used the methods to improve the segmentation accuracy include attention mechanism, multi-scale feature information fusion. Among them, U-Net achieves the best results including that the final dice coefficient of CT segmentation is 0.991 and the final dice coefficient of CTA segmentation is 0.984. In addition, attention U-Net and nnU-Net network also performs well. In this review paper, the methods chosen by 12 teams from different research groups are evaluated and their segmentation results are analyzed for the study and references to those involved.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.3233/XST-210956DOI Listing
August 2021

Prediction of acute coronary syndrome within 3 years using radiomics signature of pericoronary adipose tissue based on coronary computed tomography angiography.

Eur Radiol 2021 Aug 25. Epub 2021 Aug 25.

Department of Radiology, Shengjing Hospital of China Medical University, Shenyang, 110004, China.

Objectives: To evaluate whether radiomics signature of pericoronary adipose tissue (PCAT) based on coronary computed tomography angiography (CCTA) could improve the prediction of future acute coronary syndrome (ACS) within 3 years.

Methods: We designed a retrospective case-control study that patients with ACS (n = 90) were well matched to patients with no cardiac events (n = 1496) during 3 years follow-up, then which were randomly divided into training and test datasets with a ratio of 3:1. A total of 107 radiomics features were extracted from PCAT surrounding lesions and 14 conventional plaque characteristics were analyzed. Radiomics score, plaque score, and integrated score were respectively calculated via a linear combination of the selected features, and their performance was evaluated with discrimination, calibration, and clinical application.

Results: Radiomics score achieved superior performance in identifying patients with future ACS within 3 years in both training and test datasets (AUC = 0.826, 0.811) compared with plaque score (AUC = 0.699, 0.640), with a significant difference of AUC between two scores in the training dataset (p = 0.009); while the improvement of integrated score discriminating capability (AUC = 0.838, 0.826) was non-significant. The calibration curves of three predictive models demonstrated a good fitness respectively (all p > 0.05). Decision curve analysis suggested that integrated score added more clinical benefit than plaque score. Stratified analysis revealed that the performance of three predictive models was not affected by tube voltage, CT version, different sites of hospital.

Conclusion: CCTA-based radiomics signature of PCAT could have the potential to predict the occurrence of subsequent ACS. Radiomics-based integrated score significantly outperformed plaque score in identifying future ACS within 3 years.

Key Points: • Plaque score based on conventional plaque characteristics had certain limitations in the prediction of ACS. • Radiomics signature of PCAT surrounding plaques could have the potential to improve the predictive ability of subsequent ACS. • Radiomics-based integrated score significantly outperformed plaque score in the identification of future ACS within 3 years.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1007/s00330-021-08109-zDOI Listing
August 2021

Automated vessel segmentation in lung CT and CTA images via deep neural networks.

J Xray Sci Technol 2021 Aug 20. Epub 2021 Aug 20.

Key Laboratory of Intelligent Computing in Medical Image, Ministry of Education, Shenyang, China.

Background: The distribution of pulmonary vessels in computed tomography (CT) and computed tomography angiography (CTA) images of lung is important for diagnosing disease, formulating surgical plans and pulmonary research.

Purpose: Based on the pulmonary vascular segmentation task of International Symposium on Image Computing and Digital Medicine 2020 challenge, this paper reviews 12 different pulmonary vascular segmentation algorithms of lung CT and CTA images and then objectively evaluates and compares their performances.

Methods: First, we present the annotated reference dataset of lung CT and CTA images. A subset of the dataset consisting 7,307 slices for training and 3,888 slices for testing was made available for participants. Second, by analyzing the performance comparison of different convolutional neural networks from 12 different institutions for pulmonary vascular segmentation, the reasons for some defects and improvements are summarized. The models are mainly based on U-Net, Attention, GAN, and multi-scale fusion network. The performance is measured in terms of Dice coefficient, over segmentation ratio and under segmentation rate. Finally, we discuss several proposed methods to improve the pulmonary vessel segmentation results using deep neural networks.

Results: By comparing with the annotated ground truth from both lung CT and CTA images, most of 12 deep neural network algorithms do an admirable job in pulmonary vascular extraction and segmentation with the dice coefficients ranging from 0.70 to 0.85. The dice coefficients for the top three algorithms are about 0.80.

Conclusions: Study results show that integrating methods that consider spatial information, fuse multi-scale feature map, or have an excellent post-processing to deep neural network training and optimization process are significant for further improving the accuracy of pulmonary vascular segmentation.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.3233/XST-210955DOI Listing
August 2021

MSDS-UNet: A multi-scale deeply supervised 3D U-Net for automatic segmentation of lung tumor in CT.

Comput Med Imaging Graph 2021 09 24;92:101957. Epub 2021 Jul 24.

Amii, University of Alberta, Edmonton, Alberta, Canada.

Lung cancer is one of the most common and deadly malignant cancers. Accurate lung tumor segmentation from CT is therefore very important for correct diagnosis and treatment planning. The automated lung tumor segmentation is challenging due to the high variance in appearance and shape of the targeting tumors. To overcome the challenge, we present an effective 3D U-Net equipped with ResNet architecture and a two-pathway deep supervision mechanism to increase the network's capacity for learning richer representations of lung tumors from global and local perspectives. Extensive experiments on two real medical datasets: the lung CT dataset from Liaoning Cancer Hospital in China with 220 cases and the public dataset of TCIA with 422 cases. Our experiments demonstrate that our model achieves an average dice score (0.675), sensitivity (0.731) and F1-score (0.682) on the dataset from Liaoning Cancer Hospital, and an average dice score (0.691), sensitivity (0.746) and F1-score (0.724) on the TCIA dataset, respectively. The results demonstrate that the proposed 3D MSDS-UNet outperforms the state-of-the-art segmentation models for segmenting all scales of tumors, especially for small tumors. Moreover, we evaluated our proposed MSDS-UNet on another challenging volumetric medical image segmentation task: COVID-19 lung infection segmentation, which shows consistent improvement in the segmentation performance.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.compmedimag.2021.101957DOI Listing
September 2021

Liver vessel segmentation based on inter-scale V-Net.

Math Biosci Eng 2021 05;18(4):4327-4340

College of Marine Electrical Engineering, Dalian Maritime University, Dalian 116000, China.

Segmentation and visualization of liver vessel is a key task in preoperative planning and computer-aided diagnosis of liver diseases. Due to the irregular structure of liver vessel, accurate liver vessel segmentation is difficult. This paper proposes a method of liver vessel segmentation based on an improved V-Net network. Firstly, a dilated convolution is introduced into the network to make the network can still enlarge the receptive field without reducing down-sampling and save detailed spatial information. Secondly, a 3D deep supervision mechanism is introduced into the network to speed up the convergence of the network and help the network learn semantic features better. Finally, inter-scale dense connections are designed in the decoder of the network to prevent the loss of high-level semantic information during the decoding process and effectively integrate multi-scale feature information. The public datasets 3Dircadb were used to perform liver vessel segmentation experiments. The average dice and sensitivity of the proposed method reached 71.6 and 75.4%, respectively, which are higher than those of the original network. The experimental results show that the improved V-Net network can automatically and accurately segment labeled or even other unlabeled liver vessels from the CT images.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.3934/mbe.2021217DOI Listing
May 2021

Classification of COVID-19 Chest CT Images Based on Ensemble Deep Learning.

J Healthc Eng 2021 20;2021:5528441. Epub 2021 Apr 20.

College of Computer Science and Engineering, Northeastern University, Shenyang 110189, China.

Novel coronavirus pneumonia (NCP) has become a global pandemic disease, and computed tomography-based (CT) image analysis and recognition are one of the important tools for clinical diagnosis. In order to assist medical personnel to achieve an efficient and fast diagnosis of patients with new coronavirus pneumonia, this paper proposes an assisted diagnosis algorithm based on ensemble deep learning. The method combines the Stacked Generalization ensemble learning with the VGG16 deep learning to form a cascade classifier, and the information constituting the cascade classifier comes from multiple subsets of the training set, each of which is used to collect deviant information about the generalization behavior of the data set, such that this deviant information fills the cascade classifier. The algorithm was experimentally validated for classifying patients with novel coronavirus pneumonia, patients with common pneumonia (CP), and normal controls, and the algorithm achieved a prediction accuracy of 93.57%, sensitivity of 94.21%, specificity of 93.93%, precision of 89.40%, and F1-score of 91.74% for the three categories. The results show that the method proposed in this paper has good classification performance and can significantly improve the performance of deep neural networks for multicategory prediction tasks.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1155/2021/5528441DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC8061232PMC
May 2021

Evaluation of the mitigation efficacy of a yeast cell wall extract toward deoxynivalenol contaminated diet fed to turbot (Scophthalmus maximus).

Ecotoxicol Environ Saf 2021 Apr 13;216:112221. Epub 2021 Apr 13.

The Key Laboratory of Aquaculture Nutrition and Feed (Ministry of Agriculture), The Key Laboratory of Mariculture (Ministry of Education), Ocean University of China, Qingdao 266003, China.

Deoxynivalenol (DON) is one of the most common mycotoxins in animal feed worldwide and causes significant threats to the animal health. Increased use of plant ingredients in aquaculture feeds increased the risk of mycotoxin contamination. To evaluate the effects of dietary deoxynivalenol (DON) on growth performance, immune response and intestinal health of turbot and the mitigation efficacy of yeast cell wall extract (YCWE) toward DON, nine isonitrogenous and isolipidic diets were formulated: Diet 1 (control): No DON added; Diets 2-5 or Diets 6-9: 0.5 or 3.0 mg added DON/kg diet + 0%, 0.1%, 0.2%, or 0.4% YCWE, respectively. Results showed that Diet 6 (3 mg/kg DON, 0% YCWE) significantly decreased weight gain, specific growth rate and feed efficiency ratio of fish and reduced immunoglobulin M and complement 4 concentrations in serum. Fish fed Diet 6 presented morphological alterations, lower activity of superoxide dismutase, catalase and total antioxidant capacity but higher malondialdehyde content, lower claudin-4 and occludin expression but higher interleukin-1β expression in intestine. Besides, Diet 6 decreased the abundance of potential helpful bacteria but increased the abundance of potential pathogens in intestine. While, dietary YCWE, especially Diet 8 (3 mg/kg DON, 0.2% YCWE) and 9 (3 mg/kg DON, 0.4% YCWE), markedly improved growth performance and immune response and enhanced the intestinal health of turbot. In conclusion, dietary YCWE could mitigate the toxic effects induced by DON in turbot, and could be used as an effective strategy to control DON contamination in fish feed.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.ecoenv.2021.112221DOI Listing
April 2021

Hi-GCN: A hierarchical graph convolution network for graph embedding learning of brain network and brain disorders prediction.

Comput Biol Med 2020 12 3;127:104096. Epub 2020 Nov 3.

Amii, University of Alberta, Edmonton, Alberta, Canada.

Purpose: Recently, brain connectivity networks have been used for the classification of neurological disorder, such as Autism Spectrum Disorders (ASD) or Alzheimer's disease (AD). Network analysis provides a new way for exploring the association between brain functional deficits and the underlying structural disruption related to brain disorders. Network embedding learning that aims to automatically learn low-dimensional representations for brain networks has drawn increasing attention in recent years.

Method: In this work we build upon graph neural network in order to learn useful representations for graph classification in an end-to-end fashion. Specifically, we propose a hierarchical GCN framework (called hi-GCN) to learn the graph feature embedding while considering the network topology information and subject's association at the same time.

Results: To demonstrate the effectiveness of our approach, we evaluate the performance of the proposed method on the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset and Autism Brain Imaging Data Exchange (ABIDE) dataset. Extensive experiments on ABIDE and ADNI datasets have demonstrated competitive performance of the hi-GCN model. Specifically, we obtain an average accuracy of 73.1%/78.5% as well as AUC of 82.3%/86.5% on ABIDE/ADNI. The comprehensive experiments demonstrate that our hi-GCN is effective for graph classification with brain disorders diagnosis.

Conclusion: The proposed hi-GCN method performs the graph embedding learning from a hierarchical perspective while considering the structure in individual brain network and the subject's correlation in the global population network, which can capture the most essential embedding features to improve the classification performance of disease diagnosis. Moreover, the proposed jointly optimizing strategy also achieves faster training and easier convergence than both the hi-GCN with pre-training and two-step supervision.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.compbiomed.2020.104096DOI Listing
December 2020

Frangi based multi-scale level sets for retinal vascular segmentation.

Comput Methods Programs Biomed 2020 Dec 10;197:105752. Epub 2020 Sep 10.

Key Laboratory of Intelligent Computing in Medical Image (MIIC), Ministry of Education, Northeastern University, Shenyang, Liaoning 110169, China; Key Laboratory of Medical Image Computing (MIC), Liaoning Province, Shenyang, Liaoning 110169, China; School of Computer Science and Engineering, Northeastern University, Shenyang, Liaoning 110169, China. Electronic address:

Retinal vascular disease has always been the focus of medical attention. However, segmentation of the retinal vessels from fundus images is still an open problem due to intensity inhomogeneity in the image and thickness diversity of the retinal vessels. In this paper, we propose Frangi based multi-scale level sets to segment retinal vessels from fundus images. Vascular structures are first enhanced by the Frangi filter with local optimal scales being obtained at the same time. The enhanced image and local optimal scales are taken considered as inputs of the proposed level set models. Effectiveness of the proposed multi-scale level sets to their scale fixed versions has been evaluated using DRIVE and STARE image repositories. In addition, the proposed level set models have been tested on the DRIVE and STARE images. Experiments show that the proposed models produce segmentation accuracy at the same level with state-of-the-art methods.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.cmpb.2020.105752DOI Listing
December 2020

The Assessment of Diet Contaminated with Aflatoxin B in Juvenile Turbot () and the Evaluation of the Efficacy of Mitigation of a Yeast Cell Wall Extract.

Toxins (Basel) 2020 09 15;12(9). Epub 2020 Sep 15.

The Key Laboratory of Aquaculture Nutrition and Feed (Ministry of Agriculture), the Key Laboratory of Mariculture (Ministry of Education), Ocean University of China, Qingdao 266003, China.

This study aimed to investigate the effects of dietary AFB on growth performance, health, intestinal microbiota communities and AFB tissue residues of turbot and evaluate the mitigation efficacy of yeast cell wall extract, Mycosorb (YCWE) toward AFB contaminated dietary treatments. Nine experimental diets were formulated: Diet 1 (control): AFB free; Diets 2-5 or Diets 6-9: 20 μg AFB/kg diet or 500 μg AFB/kg diet + 0%, 0.1%, 0.2%, or 0.4% YCWE, respectively). The results showed that Diet 6 significantly decreased the concentrations of TP, GLB, C3, C4, T-CHO, TG but increased the activities of AST, ALT in serum, decreased the expressions of CAT, SOD, GPx, CYP1A but increased the expressions of CYP3A, GST-, p53 in liver. Diet 6 increased the AFB residues in serum and muscle, altered the intestinal microbiota composition, decreased the bacterial community diversity and the abundance of some potential probiotics. However, Diet 8 and Diet 9 restored the immune response, relieved adverse effects in liver, lowered the AFB residues in turbot tissues, promoted intestinal microbiota diversity and lowered the abundance of potentially pathogens. In conclusion, YCWE supplementation decreased the health effects of AFB on turbot, restoring biomarkers closer to the mycotoxin-free control diet.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.3390/toxins12090597DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7551837PMC
September 2020

Vessel segmentation using multiscale vessel enhancement and a region based level set model.

Comput Med Imaging Graph 2020 10 24;85:101783. Epub 2020 Aug 24.

Key Laboratory of Intelligent Computing in Medical Image (MIIC), Ministry of Education, Northeastern University, Shenyang, Liaoning 110169, China; Key Laboratory of Medical Image Computing (MIC), Liaoning Province, Shenyang, Liaoning 110169, China; School of Computer Science and Engineering, Northeastern University, Shenyang, Liaoning 110169, China. Electronic address:

Vessel segmentation has always been a considerable challenge task due to the presence of varying thickness of the vessels and weak contrasts of medical image intensities. In this paper, an effective method is proposed, which consists of four steps. Firstly, the input images are converted into gray ones with predetermined weightings aiming at increasing image contrast if they are colorful. Secondly, the image intensities are expanded from regions of interest to the whole image domain with a mirroring operation to avoid introducing undesired boundaries by image filtering operations existing in the next step. Thirdly, an improved multi-scale enhancement method inspired by the Frangi filtering is proposed to enhance image contrast between blood vessels and other objects in the image. Finally, an improved level set model is proposed to segment blood vessels from the enhance images and the original gray images. The proposed method has been evaluated on two retinal vessel image repositories, namely, DRIVE and STARE. Experimental results and comparison with 13 existing methods show that the proposed method produces similar segmentation accuracy at the same level with representative methods in the literature. Its effectiveness on segmentation of other type vessels is also discussed at the end of this paper.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.compmedimag.2020.101783DOI Listing
October 2020

A Global Inhomogeneous Intensity Clustering- (GINC-) Based Active Contour Model for Image Segmentation and Bias Correction.

Comput Math Methods Med 2020 1;2020:7595174. Epub 2020 Jun 1.

Key Laboratory of Medical Image Computing (MIC), Shenyang, Liaoning 110169, China.

Image segmentation is still an open problem especially when intensities of the objects of interest are overlapped due to the presence of intensity inhomogeneities. A bias correction embedded level set model is proposed in this paper where inhomogeneities are estimated by orthogonal primary functions. First, an inhomogeneous intensity clustering energy is defined based on global distribution characteristics of the image intensities, and membership functions of the clusters described by the level set function are then introduced to define the data term energy of the proposed model. Second, a regularization term and an arc length term are also included to regularize the level set function and smooth its zero-level set contour, respectively. Third, the proposed model is extended to multichannel and multiphase patterns to segment colorful images and images with multiple objects, respectively. Experimental results and comparison with relevant models demonstrate the advantages of the proposed model in terms of bias correction and segmentation accuracy on widely used synthetic and real images and the BrainWeb and the IBSR image repositories.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1155/2020/7595174DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7285411PMC
April 2021

A coronary artery segmentation method based on region growing with variable sector search area.

Technol Health Care 2020 ;28(S1):463-472

School of Computer Science and Engineering, Northeastern University, Shenyang, Liaoning, China.

Background: Coronary artery image segmentation is an important auxiliary method for coronary artery disease diagnosis.

Objective: The classical region growing algorithms, which only consider the intensity of pixels, are noise-sensitive and require manual interaction. To this end, recent methods simultaneously consider the intensity of pixels and multi-scale analysis with the region growing. Nevertheless, these methods are not fully optimized and they suffer from the drawbacks of over- or under-segmentation in many cases.

Methods: In this paper, we propose a region growing based coronary artery segmentation method. Different from the existing methods, the variable sector search area is considered in the region growing technique. A growing rule is proposed to segment the vessel, which combines the Hessian vector and the region growing with the variable sector search area. To further improve the quality of segmentation, we propose an optimization of removing some small disconnected regions.

Results: Our proposed method can search more branches while segmenting the vessel, even the small ones. It keeps an acceptable performance when dealing with stenosis and large curvature of blood vessels.

Conclusions: Quantitative evaluations are conducted on coronary angiography and the results show that the proposed method achieves a higher DSC ratio and a more reliable sensitivity ratio.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.3233/THC-209047DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7369112PMC
April 2021

A Review of the Application of Virtual Reality Technology in the Diagnosis and Treatment of Cognitive Impairment.

Front Aging Neurosci 2019 18;11:280. Epub 2019 Oct 18.

Cyberspace Institute of Advanced Technology, Guangzhou University, Guangzhou, China.

At present, with the development of an aging society and an increase in the number of elderly people, in order to ensure the ability and enthusiasm of the elderly to live independently, it is necessary to ensure that they can understand the world in a normal way. More and more elderly people have cognitive impairment, and virtual reality (VR) technology is more effective in cognitive diagnosis and treatment than traditional methods. This review article describes some studies on cognitive diagnosis and training for the elderly, and puts forward some suggestions for current studies, in the hopes that VR technology can be better applied to cognitive diagnosis and training.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.3389/fnagi.2019.00280DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC6813180PMC
October 2019

Efficient multi-kernel multi-instance learning using weakly supervised and imbalanced data for diabetic retinopathy diagnosis.

Comput Med Imaging Graph 2018 11 25;69:112-124. Epub 2018 Aug 25.

Computing Science, University of Alberta, Edmonton, Alberta, Canada.

Objective: Diabetic retinopathy (DR) is one of the most serious complications of diabetes. Early detection and treatment of DR are key public health interventions that can significantly reduce the risk of vision loss. How to effectively screen and diagnose the retinal fundus image in order to identify retinopathy in time is a major challenge. In the traditional DR screening system, the accuracy of micro-aneurysm (MA) and hemorrhagic (H) lesion detection determines the final screening performance. The detection method produced a large number of false positive samples for guaranteeing high sensitivity, and the classification model was not effective in removing false positives since the suspicious lesions lack label information.

Methods: In order to solve the problem of supervised learning in the diagnosis of DR, we formulate weakly supervised multi-class DR grading as a multi-class multi-instance problem where each image (bag) is labeled as healthy or abnormal and consists of unlabeled candidate lesion regions (instances). Specifically, we proposed a multi-kernel multi-instance learning method based on graph kernel. Moreover, we develop an under-sampling from instance level and over-sampling from bag level to improve the performance of the multi-instance learning in the diagnosis of DR.

Results: Through empirical evaluation and comparison with different baselinemethods and the state-of-the-art methods on data from Messidor, we illustrate that the proposed method reports favorable results, with an overall classification accuracy of 0.916 and an AUC of 0.957.

Conclusions: The experiments results demonstrate that the proposed multi-kernel multi-instance learning framework with bi-level re-sampling can solve the problem in the imbalanced and weakly supervised data for grading diabetic retinopathy, and it improves the diagnosis performance over several state-of-the-art competing methods.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.compmedimag.2018.08.008DOI Listing
November 2018

Generalized fused group lasso regularized multi-task feature learning for predicting cognitive outcomes in Alzheimers disease.

Comput Methods Programs Biomed 2018 Aug 3;162:19-45. Epub 2018 May 3.

Computing Science, University of Alberta, Edmonton, Alberta, Canada.

Objective: Alzheimers disease (AD) is characterized by gradual neurodegeneration and loss of brain function, especially for memory during early stages. Regression analysis has been widely applied to AD research to relate clinical and biomarker data such as predicting cognitive outcomes from Magnetic Resonance Imaging (MRI) measures. Recently, the multi-task feature learning (MTFL) methods have been widely studied to predict cognitive outcomes and select the discriminative feature subset from MRI features by incorporating inherent correlations among multiple clinical cognitive measures. However, the existing MTFL assumes the correlation among all the tasks is uniform, and the task relatedness is modeled by encouraging a common subset of features with neglecting the inherent structure of tasks and MRI features.

Methods: In this paper, we proposed a generalized fused group lasso (GFGL) regularization to model the underlying structures, involving (1) a graph structure within tasks and (2) a group structure among the image features. Then, we present a multi-task learning framework (called GFGL-MTFL), combining the ℓ-norm with the GFGL regularization, to model the flexible structures.

Results: Through empirical evaluation and comparison with different baseline methods and the state-of-the-art MTL methods on data from Alzheimer's Disease Neuroimaging Initiative (ADNI) database, we illustrate that the proposed GFGL-MTFL method outperforms other methods in terms of both Mean Squared Error (nMSE) and weighted correlation coefficient (wR). Improvements are statistically significant for most scores (tasks).

Conclusions: The experimental results with real and synthetic data demonstrate that incorporating the two prior structures by the generalized fused group lasso norm into the multi task feature learning can improve the prediction performance over several state-of-the-art competing methods, and the estimated correlation of the cognitive functions and the identification of cognition relevant imaging markers are clinically and biologically meaningful.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.cmpb.2018.04.028DOI Listing
August 2018

Linearized and Kernelized Sparse Multitask Learning for Predicting Cognitive Outcomes in Alzheimer's Disease.

Comput Math Methods Med 2018 24;2018:7429782. Epub 2018 Jan 24.

Computer Science and Engineering, Northeastern University, Shenyang, China.

Alzheimer's disease (AD) has been not only the substantial financial burden to the health care system but also the emotional burden to patients and their families. Predicting cognitive performance of subjects from their magnetic resonance imaging (MRI) measures and identifying relevant imaging biomarkers are important research topics in the study of Alzheimer's disease. Recently, the multitask learning (MTL) methods with sparsity-inducing norm (e.g., -norm) have been widely studied to select the discriminative feature subset from MRI features by incorporating inherent correlations among multiple clinical cognitive measures. However, these previous works formulate the prediction tasks as a linear regression problem. The major limitation is that they assumed a linear relationship between the MRI features and the cognitive outcomes. Some multikernel-based MTL methods have been proposed and shown better generalization ability due to the nonlinear advantage. We quantify the power of existing linear and nonlinear MTL methods by evaluating their performance on cognitive score prediction of Alzheimer's disease. Moreover, we extend the traditional -norm to a more general -norm ( ≥ 1). Experiments on the Alzheimer's Disease Neuroimaging Initiative database showed that the nonlinear -MKMTL method not only achieved better prediction performance than the state-of-the-art competitive methods but also effectively fused the multimodality data.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1155/2018/7429782DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC5830285PMC
October 2018

Nonlinearity-aware based dimensionality reduction and over-sampling for AD/MCI classification from MRI measures.

Comput Biol Med 2017 12 6;91:21-37. Epub 2017 Oct 6.

Computing Science, University of Alberta, Edmonton, Alberta, Canada.

Alzheimer's disease (AD) has been not only a substantial financial burden to the health care system but also an emotional burden to patients and their families. Making accurate diagnosis of AD based on brain magnetic resonance imaging (MRI) is becoming more and more critical and emphasized at the earliest stages. However, the high dimensionality and imbalanced data issues are two major challenges in the study of computer aided AD diagnosis. The greatest limitations of existing dimensionality reduction and over-sampling methods are that they assume a linear relationship between the MRI features (predictor) and the disease status (response). To better capture the complicated but more flexible relationship, we propose a multi-kernel based dimensionality reduction and over-sampling approaches. We combined Marginal Fisher Analysis with ℓ-norm based multi-kernel learning (MKMFA) to achieve the sparsity of region-of-interest (ROI), which leads to simultaneously selecting a subset of the relevant brain regions and learning a dimensionality transformation. Meanwhile, a multi-kernel over-sampling (MKOS) was developed to generate synthetic instances in the optimal kernel space induced by MKMFA, so as to compensate for the class imbalanced distribution. We comprehensively evaluate the proposed models for the diagnostic classification (binary class and multi-class classification) including all subjects from the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset. The experimental results not only demonstrate the proposed method has superior performance over multiple comparable methods, but also identifies relevant imaging biomarkers that are consistent with prior medical knowledge.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.compbiomed.2017.10.002DOI Listing
December 2017

The improved differential demon algorithm.

Technol Health Care 2017 Jul;25(S1):251-257

Background: Differential demon is a fast and efficient registration algorithm. It drives the floating image to deform using the force based on the gradient between the reference and floating image. But it will cause abnormal deformation when the driving force approaches zero,which limits its practical applications.

Objective: This paper proposed an improved differential demon algorithm, which aimed to enhance the registration performance of the existing demon algorithm.

Methods: Firstly, we review the original differential demon algorithm. Then, we propose an improved differential demon algorithm and the process of mathematical deduction. Finally, we use experiment to prove that the improved differential demon algorithm is effective and it can improve the accuracy of registration.

Results: We tested our method on data sets provided by Xuanwu Hospital Capital Medical University. The registration performance proved to be better than the original demon algorithm in terms of mutual information, normalized correlation coefficient, mean square error and iteration number.

Conclusions: Experiment results demonstrate the superiority of method proposed in this paper to the original demon algorithm.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.3233/THC-171328DOI Listing
July 2017

Automatic optic disc localization and segmentation in retinal images by a line operator and level sets.

Technol Health Care 2016 Apr;24 Suppl 2:S767-76

School of Information Science & Engineering, Northeastern University, Shenyang, Liaoning, China.

Background: Existing methods may fail to locate and segment the optic disc (OD) due to imprecise boundaries, inconsistent image contrast and deceptive edge features in retinal images.

Objective: To locate the OD and detect the OD boundary accurately.

Methods: The method exploits a multi-stage strategy in the detection procedure. Firstly, OD location candidate regions are identified based on high-intensity feature and vessels convergence property. Secondly, a line operator filter for circular brightness feature detection is designed to locate the OD accurately on candidates. Thirdly, an initialized contour is obtained by iterative thresholding and ellipse fitting based on the detected OD position. Finally, a region-based active contour model in a variational level set formulation and ellipse fitting are employed to estimate the OD boundary.

Results: The proposed methodology achieves an accuracy of 98.67% for OD identification and a mean distance to the closest point of 2 pixels in detecting the OD boundary.

Conclusion: The results illuminate that the proposed method is effective in the fast, automatic, and accurate localization and boundary detection of the OD. The present work contributes to the more effective evaluation of the OD and realizing automatic screening system for early eye diseases to a large extent.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.3233/THC-161206DOI Listing
April 2016

Restricted Boltzmann machines based oversampling and semi-supervised learning for false positive reduction in breast CAD.

Biomed Mater Eng 2015 ;26 Suppl 1:S1541-7

Medical Image Computing Laboratory of Ministry of Education, Northeastern University, 110819, Shenyang, China.

The false-positive reduction (FPR) is a crucial step in the computer aided detection system for the breast. The issues of imbalanced data distribution and the limitation of labeled samples complicate the classification procedure. To overcome these challenges, we propose oversampling and semi-supervised learning methods based on the restricted Boltzmann machines (RBMs) to solve the classification of imbalanced data with a few labeled samples. To evaluate the proposed method, we conducted a comprehensive performance study and compared its results with the commonly used techniques. Experiments on benchmark dataset of DDSM demonstrate the effectiveness of the RBMs based oversampling and semi-supervised learning method in terms of geometric mean (G-mean) for false positive reduction in Breast CAD.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.3233/BME-151453DOI Listing
July 2016

An image-enhancement method based on variable-order fractional differential operators.

Biomed Mater Eng 2015 ;26 Suppl 1:S1325-33

Key Laboratory of Medical Image Computing of Ministry of Education, Northeastern University, Shenyang, Liaoning 110819, China.

In this study, we develop a new algorithm based on fractional operators of variable-order in order to enhance image quality. First, three kinds of popular high-order discrete formulas are adopted to obtain the coefficients, and subsequently, a mask optimization method for selecting the fractional order adaptively is applied to construct a variable-order fractional differential mask along with the coefficients generated from the first step. We carry out experiments on OCT thoracic aorta images and some nature images with low contrast and noise, demonstrating that the high-order discrete method leads to significantly better performance in enhancing the edge information nonlinearly compared to the standard first-order discrete method. Moreover, the optimized mask with variable-order of the fractional derivative not only can preserve the edge information of the processed images adequately, but it also effectively suppresses the noise in the smooth area.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.3233/BME-151430DOI Listing
July 2016

Improved Hessian multiscale enhancement filter.

Biomed Mater Eng 2014 ;24(6):3267-75

Key Laboratory of Medical Image Computing of Northeastern University, Ministry of Education, Shenyang 110819, China College of Information Science and Engineering, Northeastern University, Shenyang 110004, China.

Traditional Hessian multiscale filter consider only the local geometric feature but not the global grayscale information. In medical image analysis, Hessian filter is usually used to enhance the blood vessels. However, it also produces some pseudo vascular structures or some isolate noise points, such as the nasal soft tissues that have the similar shape with the vessels in MRA data, which will increase the difficulty of cerebrovascular segmentation. To resolve this issue, an improved Hessian multiscale filter is proposed in this paper. An image grayscale factor is added to the vascular similarity function computed by Hessian matrix eigenvalue. This method is experimented on brain MRA data and lung CTA data. Experimental results show that this method can enhance vascular structures, and simultaneously reduce the appearance of the pseudo vascular structures and the isolated noise points.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.3233/BME-141149DOI Listing
June 2015

Accurate airway centerline extraction based on topological thinning using graph-theoretic analysis.

Biomed Mater Eng 2014 ;24(6):3239-49

Key Laboratory of Medical Image Computing of Ministry of Education, Northeastern University, No. 3-11, Wenhua Road, Heping District, Shenyang, P.R. China College of Information Science and Engineering, Northeastern University, No. 3-11, Wenhua Road, Heping District, Shenyang, P.R. China.

The quantitative analysis of the airway tree is of critical importance in the CT-based diagnosis and treatment of popular pulmonary diseases. The extraction of airway centerline is a precursor to identify airway hierarchical structure, measure geometrical parameters, and guide visualized detection. Traditional methods suffer from extra branches and circles due to incomplete segmentation results, which induce false analysis in applications. This paper proposed an automatic and robust centerline extraction method for airway tree. First, the centerline is located based on the topological thinning method; border voxels are deleted symmetrically to preserve topological and geometrical properties iteratively. Second, the structural information is generated using graph-theoretic analysis. Then inaccurate circles are removed with a distance weighting strategy, and extra branches are pruned according to clinical anatomic knowledge. The centerline region without false appendices is eventually determined after the described phases. Experimental results show that the proposed method identifies more than 96% branches and keep consistency across different cases and achieves superior circle-free structure and centrality.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.3233/BME-141146DOI Listing
June 2015

Ensemble-based hybrid probabilistic sampling for imbalanced data learning in lung nodule CAD.

Comput Med Imaging Graph 2014 Apr 21;38(3):137-50. Epub 2013 Dec 21.

Computing Science, University of Alberta, Edmonton, Alberta, Canada. Electronic address:

Classification plays a critical role in false positive reduction (FPR) in lung nodule computer aided detection (CAD). The difficulty of FPR lies in the variation of the appearances of the nodules, and the imbalance distribution between the nodule and non-nodule class. Moreover, the presence of inherent complex structures in data distribution, such as within-class imbalance and high-dimensionality are other critical factors of decreasing classification performance. To solve these challenges, we proposed a hybrid probabilistic sampling combined with diverse random subspace ensemble. Experimental results demonstrate the effectiveness of the proposed method in terms of geometric mean (G-mean) and area under the ROC curve (AUC) compared with commonly used methods.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.compmedimag.2013.12.003DOI Listing
April 2014

Adaptive fast marching method for automatic liver segmentation from CT images.

Med Phys 2013 Sep;40(9):091917

College of Computer and Information Technology, Nanyang Normal University, Nanyang 473061, China.

Purpose: Liver segmentation is a fundamental step in computer-aided liver disease diagnosis and surgery planning. For the sake of high accuracy and efficiency, in this study, the authors present an automatic seed point selection method and an adaptive fast marching method (FMM) for liver segmentation.

Methods: The automatic seed point selection method is according to the structure and intensity characteristics of liver. The proposed adaptive FMM is self-adaptive parameter adjustment. The arrival time parameter T in FMM is adjusted according to the intensity statistics of the possible liver region, which can be used to estimate the size of liver region on the corresponding computed tomography (CT) slices. The proposed algorithm consists of the following steps. First, a thresholding operation was applied to remove the ribs, spines, and kidneys, followed by a smooth filter for noise reduction and a nonlinear gray scale converter, which was used to enhance the contrast of the liver parenchyma. Second, the seed points located in the liver were selected automatically. Finally, using the processed image as a speed function, adaptive FMM was employed to generate the liver contour.

Results: Clinical validation has been performed on 30 abdominal CT data-sets. The proposed algorithm achieved an overall true positive rate of 98.7%, false negative rate of 1.6%, false positive rate of 5.2%, and the DICE coefficient of 96.7%. It takes about 0.30s for a 512 × 512-pixel slice.

Conclusions: The method has been applied successfully to achieve fast and accurate liver segmentation.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1118/1.4819824DOI Listing
September 2013

A distance-field based automatic neuron tracing method.

BMC Bioinformatics 2013 Mar 12;14:93. Epub 2013 Mar 12.

Janelia Farm Research Campus, Howard Hughes Medical Institute, Ashburn, VA, USA.

Background: Automatic 3D digital reconstruction (tracing) of neurons embedded in noisy microscopic images is challenging, especially when the cell morphology is complex.

Results: We have developed a novel approach, named DF-Tracing, to tackle this challenge. This method first extracts the neurite signal (foreground) from a noisy image by using anisotropic filtering and automated thresholding. Then, DF-Tracing executes a coupled distance-field (DF) algorithm on the extracted foreground neurite signal and reconstructs the neuron morphology automatically. Two distance-transform based "force" fields are used: one for "pressure", which is the distance transform field of foreground pixels (voxels) to the background, and another for "thrust", which is the distance transform field of the foreground pixels to an automatically determined seed point. The coupling of these two force fields can "push" a "rolling ball" quickly along the skeleton of a neuron, reconstructing the 3D cell morphology.

Conclusion: We have used DF-Tracing to reconstruct the intricate neuron structures found in noisy image stacks, obtained with 3D laser microscopy, of dragonfly thoracic ganglia. Compared to several previous methods, DF-Tracing produces better reconstructions.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1186/1471-2105-14-93DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3637550PMC
March 2013

Eight pairs of descending visual neurons in the dragonfly give wing motor centers accurate population vector of prey direction.

Proc Natl Acad Sci U S A 2013 Jan 3;110(2):696-701. Epub 2012 Dec 3.

Janelia Farm Research Campus, Howard Hughes Medical Institute, Ashburn, VA 20147, USA.

Intercepting a moving object requires prediction of its future location. This complex task has been solved by dragonflies, who intercept their prey in midair with a 95% success rate. In this study, we show that a group of 16 neurons, called target-selective descending neurons (TSDNs), code a population vector that reflects the direction of the target with high accuracy and reliability across 360°. The TSDN spatial (receptive field) and temporal (latency) properties matched the area of the retina where the prey is focused and the reaction time, respectively, during predatory flights. The directional tuning curves and morphological traits (3D tracings) for each TSDN type were consistent among animals, but spike rates were not. Our results emphasize that a successful neural circuit for target tracking and interception can be achieved with few neurons and that in dragonflies this information is relayed from the brain to the wing motor centers in population vector form.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1073/pnas.1210489109DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3545807PMC
January 2013

White matter abnormalities in children and adolescents with temporal lobe epilepsy.

Magn Reson Imaging 2010 Nov 24;28(9):1290-8. Epub 2010 Jul 24.

MEG Center, Cincinnati Children's Hospital Medical Center, Cincinnati, OH 45220, USA.

Background And Purpose: The widespread propagation of synchronized neuronal firing in seizure disorders may affect cortical and subcortical brain regions. Diffusion tensor imaging (DTI) can noninvasively quantify white matter integrity. The purpose of this study was to investigate the abnormal changes of white matter in children and adolescents with focal temporal lobe epilepsy (TLE) using DTI.

Materials And Methods: Eight patients with clinically diagnosed TLE and eight age- and sex-matched healthy controls were studied. DTI images were obtained with a 3-T magnetic resonance imaging scanner. The epileptic foci were localized with magnetoencephalography. Fractional anisotropy (FA), mean diffusivity (MD), parallel (λ(||)) and perpendicular (λ(⊥)) diffusivities in the genu of the corpus callosum, splenium of the corpus callosum (SCC), external capsule (EC), anterior limbs of the internal capsule (AIC), and the posterior limbs of the internal capsule (PIC) were calculated. The DTI parameters between patients and controls were statistically compared. Correlations of these DTI parameters of each selected structure with age of seizure onset and duration of epilepsy were analysed.

Results: In comparison to controls, both patients' seizure ipsilateral and contralateral had significantly lower FA in the AIC; PIC and SCC and higher MD, λ(||) and λ(⊥) in the EC, AIC, PIC and SCC. The MD, λ(||) and λ(⊥) were significantly correlated with age of seizure onset in the EC and PIC. λ(||) was significantly correlated with the duration of epilepsy in the EC and PIC.

Conclusion: The results of the present study indicate that children and adolescents with TLE had significant abnormalities in the white matter in the hemisphere with seizure foci. Furthermore, these abnormalities may extend to the other brain hemisphere. The age of seizure onset and duration of epilepsy may be important factors in determining the extent of influence of children and adolescents TLE on white matter.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.mri.2010.03.046DOI Listing
November 2010
-->