Pub Date : 2014-07-31DOI: 10.1109/ISBI.2014.6867901
Weijia Chen, L. Itu, Puneet S. Sharma, A. Kamen
In this paper, we present a framework for uncertainty quantification in medical image-based patient-specific hemodynamic computations. To illustrate the overall methodology, we have used an aortic coarctation model for computing trans-stenotic pressure gradient. Variance-based Sobol sensitivity indices are used to evaluate the relative influence of the various uncertain measurements and model parameters on the global variance of the output. Next, a generalized Polynomial Chaos Expansion (PCE) method is used to quantify the uncertainties in the computed mean and peak pressure gradient in terms of a probability density functions and error bars over a full cardiac cycle.
{"title":"Uncertainty quantification in medical image-based hemodynamic computations","authors":"Weijia Chen, L. Itu, Puneet S. Sharma, A. Kamen","doi":"10.1109/ISBI.2014.6867901","DOIUrl":"https://doi.org/10.1109/ISBI.2014.6867901","url":null,"abstract":"In this paper, we present a framework for uncertainty quantification in medical image-based patient-specific hemodynamic computations. To illustrate the overall methodology, we have used an aortic coarctation model for computing trans-stenotic pressure gradient. Variance-based Sobol sensitivity indices are used to evaluate the relative influence of the various uncertain measurements and model parameters on the global variance of the output. Next, a generalized Polynomial Chaos Expansion (PCE) method is used to quantify the uncertainties in the computed mean and peak pressure gradient in terms of a probability density functions and error bars over a full cardiac cycle.","PeriodicalId":440405,"journal":{"name":"2014 IEEE 11th International Symposium on Biomedical Imaging (ISBI)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125516546","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-07-31DOI: 10.1109/ISBI.2014.6868037
J. Tarquino, A. Rueda, E. Romero
Spatial resolution of Diffusion Weighted (DW) images is currently limited by diverse considerations. This situation introduces a series of artifacts, such as the partial volume effect (PVE), that therefore affect the sensitivity of DW imaging analysis. In this paper, a new multiscale/sparse super-resolution method increases the spatial resolution of the DW images. Based on the redundancy presented in this kind of images, the proposed method uses local information and the multiscale shearlet transformation to closely approach the DW image acquisition process. A comparison of this proposal with a classical image interpolation method demonstrates an improvement of 2.27 dB in the PSNR measure and 1.67% in the SSIM metric.
{"title":"A multiscale/sparse representation for Diffusion Weighted Imaging (DWI) super-resolution","authors":"J. Tarquino, A. Rueda, E. Romero","doi":"10.1109/ISBI.2014.6868037","DOIUrl":"https://doi.org/10.1109/ISBI.2014.6868037","url":null,"abstract":"Spatial resolution of Diffusion Weighted (DW) images is currently limited by diverse considerations. This situation introduces a series of artifacts, such as the partial volume effect (PVE), that therefore affect the sensitivity of DW imaging analysis. In this paper, a new multiscale/sparse super-resolution method increases the spatial resolution of the DW images. Based on the redundancy presented in this kind of images, the proposed method uses local information and the multiscale shearlet transformation to closely approach the DW image acquisition process. A comparison of this proposal with a classical image interpolation method demonstrates an improvement of 2.27 dB in the PSNR measure and 1.67% in the SSIM metric.","PeriodicalId":440405,"journal":{"name":"2014 IEEE 11th International Symposium on Biomedical Imaging (ISBI)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116388013","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-07-31DOI: 10.1109/ISBI.2014.6868045
Siqi Liu, Sidong Liu, Weidong (Tom) Cai, Sonia Pujol, R. Kikinis, D. Feng
The accurate diagnosis of Alzheimer's disease (AD) plays a significant role in patient care, especially at the early stage, because the consciousness of the severity and the progression risks allows the patients to take prevention measures before irreversible brain damages are shaped. Although many studies have applied machine learning methods for computer-aided-diagnosis (CAD) of AD recently, a bottleneck of the diagnosis performance was shown in most of the existing researches, mainly due to the congenital limitations of the chosen learning models. In this study, we design a deep learning architecture, which contains stacked auto-encoders and a softmax output layer, to overcome the bottleneck and aid the diagnosis of AD and its prodromal stage, Mild Cognitive Impairment (MCI). Compared to the previous workflows, our method is capable of analyzing multiple classes in one setting, and requires less labeled training samples and minimal domain prior knowledge. A significant performance gain on classification of all diagnosis groups was achieved in our experiments.
{"title":"Early diagnosis of Alzheimer's disease with deep learning","authors":"Siqi Liu, Sidong Liu, Weidong (Tom) Cai, Sonia Pujol, R. Kikinis, D. Feng","doi":"10.1109/ISBI.2014.6868045","DOIUrl":"https://doi.org/10.1109/ISBI.2014.6868045","url":null,"abstract":"The accurate diagnosis of Alzheimer's disease (AD) plays a significant role in patient care, especially at the early stage, because the consciousness of the severity and the progression risks allows the patients to take prevention measures before irreversible brain damages are shaped. Although many studies have applied machine learning methods for computer-aided-diagnosis (CAD) of AD recently, a bottleneck of the diagnosis performance was shown in most of the existing researches, mainly due to the congenital limitations of the chosen learning models. In this study, we design a deep learning architecture, which contains stacked auto-encoders and a softmax output layer, to overcome the bottleneck and aid the diagnosis of AD and its prodromal stage, Mild Cognitive Impairment (MCI). Compared to the previous workflows, our method is capable of analyzing multiple classes in one setting, and requires less labeled training samples and minimal domain prior knowledge. A significant performance gain on classification of all diagnosis groups was achieved in our experiments.","PeriodicalId":440405,"journal":{"name":"2014 IEEE 11th International Symposium on Biomedical Imaging (ISBI)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122665384","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-07-31DOI: 10.1109/ISBI.2014.6867927
Hongming Li, Yong Fan
A novel method is proposed to parcellate the cerebral cortex into functionally homogenous regions at multiple scales with a hierarchical organization based on resting-state fMRI data. The cortical vertices are clustered according to inter-vertex functional similarity measures progressively at multiple spatial scales from fine to coarse by a procedure referred to as floating aggregation. The floating aggregation takes into consideration both the inter-regional functional similarity and the consistency of intra-regional functional homogeneity measures at every level of the resulting parcellation hierarchy. This aggregation procedure does not require to specify the number of regions for the parcellation, and could help identify proper spatial scales for the brain parcellation based on the overall region homogeneity changes across levels of the hierarchy. The experimental results on a resting-state fMRI dataset have demonstrated that the proposed method could not only obtain brain parcellation results with better functional homogeneity measures than state-of-the-art techniques, but also identify a hierarchical functional organization of the brain at multiple spatial scales.
{"title":"Hierarchical organization of the functional brain identified using floating aggregation of functional signals","authors":"Hongming Li, Yong Fan","doi":"10.1109/ISBI.2014.6867927","DOIUrl":"https://doi.org/10.1109/ISBI.2014.6867927","url":null,"abstract":"A novel method is proposed to parcellate the cerebral cortex into functionally homogenous regions at multiple scales with a hierarchical organization based on resting-state fMRI data. The cortical vertices are clustered according to inter-vertex functional similarity measures progressively at multiple spatial scales from fine to coarse by a procedure referred to as floating aggregation. The floating aggregation takes into consideration both the inter-regional functional similarity and the consistency of intra-regional functional homogeneity measures at every level of the resulting parcellation hierarchy. This aggregation procedure does not require to specify the number of regions for the parcellation, and could help identify proper spatial scales for the brain parcellation based on the overall region homogeneity changes across levels of the hierarchy. The experimental results on a resting-state fMRI dataset have demonstrated that the proposed method could not only obtain brain parcellation results with better functional homogeneity measures than state-of-the-art techniques, but also identify a hierarchical functional organization of the brain at multiple spatial scales.","PeriodicalId":440405,"journal":{"name":"2014 IEEE 11th International Symposium on Biomedical Imaging (ISBI)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122889136","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-07-31DOI: 10.1109/ISBI.2014.6867921
Zhaoxia Zhang, Ming Yan, Kun Tao, Xiao Xuan
In tomosynthesis imaging, two kinds of metal artifacts will influence diagnosis: undershooting and ripple. In this paper we propose a novel metal artifact reduction (MAR) algorithm to reduce the both of these effects. First, the raw projection data are analyzed and metal areas are identified through segmentation. Then the metal areas are filled with an interpolated value based on the neighborhood background (non-segmented) pixels. The filled regions and metal regions are then reconstructed separately with Filtered Backprojection(FBP). Lastly, the two reconstruction results are combined together to get the final artifacts-free images. Phantom and clinical images are evaluated using qualitative and quantitative methods which demonstrate the algorithms effectiveness.
{"title":"Metal artifacts reduction for tomosynthesis","authors":"Zhaoxia Zhang, Ming Yan, Kun Tao, Xiao Xuan","doi":"10.1109/ISBI.2014.6867921","DOIUrl":"https://doi.org/10.1109/ISBI.2014.6867921","url":null,"abstract":"In tomosynthesis imaging, two kinds of metal artifacts will influence diagnosis: undershooting and ripple. In this paper we propose a novel metal artifact reduction (MAR) algorithm to reduce the both of these effects. First, the raw projection data are analyzed and metal areas are identified through segmentation. Then the metal areas are filled with an interpolated value based on the neighborhood background (non-segmented) pixels. The filled regions and metal regions are then reconstructed separately with Filtered Backprojection(FBP). Lastly, the two reconstruction results are combined together to get the final artifacts-free images. Phantom and clinical images are evaluated using qualitative and quantitative methods which demonstrate the algorithms effectiveness.","PeriodicalId":440405,"journal":{"name":"2014 IEEE 11th International Symposium on Biomedical Imaging (ISBI)","volume":"74 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122146921","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-07-31DOI: 10.1109/ISBI.2014.6867879
M. Berger, Klaus Sembritzki, J. Hornegger, Christina Bauer
In the last decade many approaches have been introduced that allow for automatic classification of brain tumors by means of pattern recognition and magnetic resonance spectroscopy. Despite promising classification accuracies, none of these methods has found its way into clinical practice, which is also related to the missing transparency for the basis of their decision making. In this work, we develop two methods to increase the interpretability of such classification systems. First we propose a new reliability measure that determines a lower bound for the probability that a particular classification is correct. Additionally, we present a method that visualizes important regions for the classifier directly in the spectral domain. As a basis for this, seven classification methods were evaluated for their performance in discriminating aggressive tumors, low-grade glioma and meningioma, based on a common database. Our results show that the novel reliability measure is in good agreement with the actual classification accuracy. Further we point out that our visualization method clearly indicates which spectral regions are important for a classifier and how metabolite concentrations correspond to specific tumor types. Combining both methods can help to better understand a classifier's decision and therefore make the outcome more transparent and trustworthy.
{"title":"Increasing the credibility of MR spectroscopy-based automatic brain tumor classification systems","authors":"M. Berger, Klaus Sembritzki, J. Hornegger, Christina Bauer","doi":"10.1109/ISBI.2014.6867879","DOIUrl":"https://doi.org/10.1109/ISBI.2014.6867879","url":null,"abstract":"In the last decade many approaches have been introduced that allow for automatic classification of brain tumors by means of pattern recognition and magnetic resonance spectroscopy. Despite promising classification accuracies, none of these methods has found its way into clinical practice, which is also related to the missing transparency for the basis of their decision making. In this work, we develop two methods to increase the interpretability of such classification systems. First we propose a new reliability measure that determines a lower bound for the probability that a particular classification is correct. Additionally, we present a method that visualizes important regions for the classifier directly in the spectral domain. As a basis for this, seven classification methods were evaluated for their performance in discriminating aggressive tumors, low-grade glioma and meningioma, based on a common database. Our results show that the novel reliability measure is in good agreement with the actual classification accuracy. Further we point out that our visualization method clearly indicates which spectral regions are important for a classifier and how metabolite concentrations correspond to specific tumor types. Combining both methods can help to better understand a classifier's decision and therefore make the outcome more transparent and trustworthy.","PeriodicalId":440405,"journal":{"name":"2014 IEEE 11th International Symposium on Biomedical Imaging (ISBI)","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128554575","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-07-31DOI: 10.1109/ISBI.2014.6867816
H. Nosato, H. Sakanashi, E. Takahashi, M. Murakawa
This study aims to establish a new method of objective evaluation for optical colonoscopy that can quantify the severity of colonic mucosa for ulcerative colitis (UC). UC is an intractable disease and has been the subject of survey research for long time. However, because there are enormous variations in the patterns of symptoms associated with UC, universal diagnostic standards have yet to be established. Accordingly, diagnostic accuracy is highly dependent on the experience and knowledge of the medical doctor. In order to overcome this problem, this paper describes a method of objective evaluations for UC based on image recognition techniques and multi-discriminant analysis. The proposed method extracts geometrical features using higher order local auto-correlations from the saturation element of the HSV color space for the colonoscopy images, and makes classifications according to the UC severity based on the subspace method. This study provides an index for UC severity to support colonoscopy diagnosis.
{"title":"An objective evaluation method of ulcerative colitis with optical colonoscopy images based on higher order local auto-correlation features","authors":"H. Nosato, H. Sakanashi, E. Takahashi, M. Murakawa","doi":"10.1109/ISBI.2014.6867816","DOIUrl":"https://doi.org/10.1109/ISBI.2014.6867816","url":null,"abstract":"This study aims to establish a new method of objective evaluation for optical colonoscopy that can quantify the severity of colonic mucosa for ulcerative colitis (UC). UC is an intractable disease and has been the subject of survey research for long time. However, because there are enormous variations in the patterns of symptoms associated with UC, universal diagnostic standards have yet to be established. Accordingly, diagnostic accuracy is highly dependent on the experience and knowledge of the medical doctor. In order to overcome this problem, this paper describes a method of objective evaluations for UC based on image recognition techniques and multi-discriminant analysis. The proposed method extracts geometrical features using higher order local auto-correlations from the saturation element of the HSV color space for the colonoscopy images, and makes classifications according to the UC severity based on the subspace method. This study provides an index for UC severity to support colonoscopy diagnosis.","PeriodicalId":440405,"journal":{"name":"2014 IEEE 11th International Symposium on Biomedical Imaging (ISBI)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128583121","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-07-31DOI: 10.1109/ISBI.2014.6868036
D. Dong, Hui Hui, Caiyun Yang, Jin Guo, Yujie Yang, Liangliang Shi, W. Mu, Jie Tian
We have designed a multimodality molecular imaging system for small animals. The aim is to develop a system which can perform functional imaging, structural imaging, and molecular imaging. Our multimodality system contains five imaging modalities which are Bioluminescence Tomography (BLT), Fluorescence Molecular Tomography (FMT), Cerenkov Luminescence Tomography (CLT), X-ray Computed Tomography (CT), and Positron Emission Tomography (PET). We have designed both the hardware structure and software to make sure multimodality imaging can be achieved. Here we will report the overall design and work flow of the system.
{"title":"Preliminary design of a multimodality molecular imaging system","authors":"D. Dong, Hui Hui, Caiyun Yang, Jin Guo, Yujie Yang, Liangliang Shi, W. Mu, Jie Tian","doi":"10.1109/ISBI.2014.6868036","DOIUrl":"https://doi.org/10.1109/ISBI.2014.6868036","url":null,"abstract":"We have designed a multimodality molecular imaging system for small animals. The aim is to develop a system which can perform functional imaging, structural imaging, and molecular imaging. Our multimodality system contains five imaging modalities which are Bioluminescence Tomography (BLT), Fluorescence Molecular Tomography (FMT), Cerenkov Luminescence Tomography (CLT), X-ray Computed Tomography (CT), and Positron Emission Tomography (PET). We have designed both the hardware structure and software to make sure multimodality imaging can be achieved. Here we will report the overall design and work flow of the system.","PeriodicalId":440405,"journal":{"name":"2014 IEEE 11th International Symposium on Biomedical Imaging (ISBI)","volume":"104 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129054505","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
We present a method to accurately localize multiple small fluorescent objects within the tissue using fluorescence molecular tomography (FMT). The proposed method exploits the localized or sparse nature of the fluorophores in the tissue as a priori information to considerably improve the accuracy of the reconstruction of fluorophore distribution. This is accomplished by minimizing a cost function that includes the L1/2 norm of the fluorophore distribution vector. To deal with the nonconvex penalty, the L1/2 regularizer is transformed into a reweighted L1-norm minimization problem and then it is efficiently solved by a homotopy-based algorithm. Simulation experiments on a 3D digital mouse atlas are performed to verify the feasibility of the proposed method, and the results demonstrate L1/2 regularization is a promising approach for image reconstruction problem of FMT.
{"title":"L1/2 regularization method for multiple-target reconstruction in fluorescent molecular tomography","authors":"Xiaowei He, Hongbo Guo, Yuqing Hou, Jingjing Yu, Hejuan Liu, Hai Zhang","doi":"10.1109/ISBI.2014.6867830","DOIUrl":"https://doi.org/10.1109/ISBI.2014.6867830","url":null,"abstract":"We present a method to accurately localize multiple small fluorescent objects within the tissue using fluorescence molecular tomography (FMT). The proposed method exploits the localized or sparse nature of the fluorophores in the tissue as a priori information to considerably improve the accuracy of the reconstruction of fluorophore distribution. This is accomplished by minimizing a cost function that includes the L1/2 norm of the fluorophore distribution vector. To deal with the nonconvex penalty, the L1/2 regularizer is transformed into a reweighted L1-norm minimization problem and then it is efficiently solved by a homotopy-based algorithm. Simulation experiments on a 3D digital mouse atlas are performed to verify the feasibility of the proposed method, and the results demonstrate L1/2 regularization is a promising approach for image reconstruction problem of FMT.","PeriodicalId":440405,"journal":{"name":"2014 IEEE 11th International Symposium on Biomedical Imaging (ISBI)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130580922","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-07-31DOI: 10.1109/ISBI.2014.6867913
F. Carpinteiro, Pedro Costa, Mario Sáenz Espinoza, Ivo M. Silva, J. Cunha
Automated tracking of axonal neurotransmitter vesicles is a challenging problem in neuroscience. The present vesicle tracking is typically performed manually over confocal microscopy images. NeuronDynamics is a method designed to automate and speed-up the characterization of global vesicle movement in neurons while yielding high accuracy and precision results (similar or better than expert clinicians). For a set of fluorescent-marked vesicles “films”, Neuron-Dynamics performs a two stage approach: 1) Training: the system asks the user to mark a set of vesicles and the position of the cellular body; 2) Detection & tracking: based on the previous training, the system runs a Bayesian classifier over the image sequence to detect and classify vesicles and their movements (speed and direction). The obtained results were compared to another state-of-the-art method (FluoTracker), and were found greatly higher in accuracy, sensitivity, specificity and precision. Although NeuronDynamics is a semi-automated process, it is significantly faster than manual tracking and can be adapted to be used for similar approaches for other biological samples.
{"title":"Neurondynamics: A method for neurotransmitter vesicle movement characterization in neurons","authors":"F. Carpinteiro, Pedro Costa, Mario Sáenz Espinoza, Ivo M. Silva, J. Cunha","doi":"10.1109/ISBI.2014.6867913","DOIUrl":"https://doi.org/10.1109/ISBI.2014.6867913","url":null,"abstract":"Automated tracking of axonal neurotransmitter vesicles is a challenging problem in neuroscience. The present vesicle tracking is typically performed manually over confocal microscopy images. NeuronDynamics is a method designed to automate and speed-up the characterization of global vesicle movement in neurons while yielding high accuracy and precision results (similar or better than expert clinicians). For a set of fluorescent-marked vesicles “films”, Neuron-Dynamics performs a two stage approach: 1) Training: the system asks the user to mark a set of vesicles and the position of the cellular body; 2) Detection & tracking: based on the previous training, the system runs a Bayesian classifier over the image sequence to detect and classify vesicles and their movements (speed and direction). The obtained results were compared to another state-of-the-art method (FluoTracker), and were found greatly higher in accuracy, sensitivity, specificity and precision. Although NeuronDynamics is a semi-automated process, it is significantly faster than manual tracking and can be adapted to be used for similar approaches for other biological samples.","PeriodicalId":440405,"journal":{"name":"2014 IEEE 11th International Symposium on Biomedical Imaging (ISBI)","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124171791","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：