| 1. |
- Fu, Keren, 1988, et al.
(författare)
-
Adaptive Multi-Level Region Merging for Salient Object Detection
- 2014
-
Ingår i: British Machine Vision Conference (BMVC) 2014. ; , s. 11 -
-
Konferensbidrag (refereegranskat)abstract
- Most existing salient object detection algorithms face the problem of either under or over-segmenting an image. More recent methods address the problem via multi-level segmentation. However, the number of segmentation levels is manually predetermined and only works well on specific class of images. In this paper, a new salient object detection scheme is presented based on adaptive multi-level region merging. A graph based merging scheme is developed to reassemble regions based on their shared contourstrength. This merging process is adaptive to complete contours of salient objects that can then be used for global perceptual analysis, e.g., foreground/ground separation. Such contour completion is enhanced by graph-based spectral decomposition. We show that even though simple region saliency measurements are adopted for each region, encouraging performance can be obtained after across-level integration. Experiments by comparing with 13 existing methods on three benchmark datasets including MSRA-1000, SOD and SED show the proposed method results in uniform object enhancement and achieves state-of-the-art performance.
|
|
| 2. |
- Fu, Keren, 1988, et al.
(författare)
-
One-class support vector machine-assisted robust tracking
- 2013
-
Ingår i: Journal of Electronic Imaging. - 1017-9909. ; 22:2, s. 11-
-
Tidskriftsartikel (refereegranskat)abstract
- Recently, tracking is regarded as a binary classification problem by discriminative tracking methods. However, such binary classification may not fully handle the outliers, which may cause drifting. We argue that tracking may be regarded as one-class problem, which avoids gathering limited negative samples for background description. Inspired by the fact the positive feature space generatedby one-class support vector machine (SVM) is bounded by a closed hyper sphere, we propose a tracking method utilizing one-class SVMs that adopt histograms of oriented gradient and 2bit binary patterns as features. Thus, it is called the one-class SVM tracker (OCST). Simultaneously, an efficient initialization and online updating scheme is proposed. Extensive experimental results prove that OCST outperforms some state-of-the-art discriminative tracking methods that tackle the problem using binary classifiers on providing accurate tracking and alleviating serious drifting.
|
|
| 3. |
- Fu, Keren, 1988, et al.
(författare)
-
One-Class SVM Assisted Accurate Tracking
- 2012
-
Ingår i: 6th ACM/IEEE Int'l Conf on Distributed Smart Cameras (ICDSC 12), Oct 30 - Nov.2, 2012, Hong Kong. ; , s. 6 pages-
-
Konferensbidrag (refereegranskat)abstract
- Recently, tracking is regarded as a binary classification problem by discriminative tracking methods. However, such binary classification may not fully handle the outliers, which may cause drifting. In this paper, we argue that tracking may be regarded as one-class problem, which avoids gathering limited negative samples for background description. Inspired by the fact the positive feature space generated by One-Class SVM is bounded by a closed sphere, we propose a novel tracking method utilizing One-Class SVMs that adopt HOG and 2 bit-BP as features, called One-Class SVM Tracker (OCST). Simultaneously an efficient initialization and online updating scheme is alsoproposed. Extensive experimental results prove that OCST outperforms some state-of-the-art discriminative tracking methods on providing accurate tracking and alleviating serious drifting.
|
|
| 4. |
- Ali, Muhaddisa Barat, 1986, et al.
(författare)
-
A Feasibility Study on Deep Learning Based Brain Tumor Segmentation Using 2D Ellipse Box Areas.
- 2022
-
Ingår i: Sensors (Basel, Switzerland). - : MDPI AG. - 1424-8220. ; 22:14
-
Tidskriftsartikel (refereegranskat)abstract
- In most deep learning-based brain tumor segmentation methods, training the deep network requires annotated tumor areas. However, accurate tumor annotation puts high demands on medical personnel. The aim of this study is to train a deep network for segmentation by using ellipse box areas surrounding the tumors. In the proposed method, the deep network is trained by using a large number of unannotated tumor images with foreground (FG) and background (BG) ellipse box areas surrounding the tumor and background, and a small number of patients (<20) with annotated tumors. The training is conducted by initial training on two ellipse boxes on unannotated MRIs, followed by refined training on a small number of annotated MRIs. We use a multi-stream U-Net for conducting our experiments, which is an extension of the conventional U-Net. This enables the use of complementary information from multi-modality (e.g., T1, T1ce, T2, and FLAIR) MRIs. To test the feasibility of the proposed approach, experiments and evaluation were conducted on two datasets for glioma segmentation. Segmentation performance on the test sets is then compared with those used on the same network but trained entirely by annotated MRIs. Our experiments show that the proposed method has obtained good tumor segmentation results on the test sets, wherein the dice score on tumor areas is (0.8407, 0.9104), and segmentation accuracy on tumor areas is (83.88%, 88.47%) for the MICCAI BraTS'17 and US datasets, respectively. Comparing the segmented results by using the network trained by all annotated tumors, the drop in the segmentation performance from the proposed approach is (0.0594, 0.0159) in the dice score, and (8.78%, 2.61%) in segmented tumor accuracy for MICCAI and US test sets, which is relatively small. Our case studies have demonstrated that training the network for segmentation by using ellipse box areas in place of all annotated tumors is feasible, and can be considered as an alternative, which is a trade-off between saving medical experts' time annotating tumors and a small drop in segmentation performance.
|
|
| 5. |
- Ali, Muhaddisa Barat, 1986, et al.
(författare)
-
A novel federated deep learning scheme for glioma and its subtype classification
- 2023
-
Ingår i: Frontiers in Neuroscience. - 1662-4548 .- 1662-453X. ; 17
-
Tidskriftsartikel (refereegranskat)abstract
- Background: Deep learning (DL) has shown promising results in molecular-based classification of glioma subtypes from MR images. DL requires a large number of training data for achieving good generalization performance. Since brain tumor datasets are usually small in size, combination of such datasets from different hospitals are needed. Data privacy issue from hospitals often poses a constraint on such a practice. Federated learning (FL) has gained much attention lately as it trains a central DL model without requiring data sharing from different hospitals. Method: We propose a novel 3D FL scheme for glioma and its molecular subtype classification. In the scheme, a slice-based DL classifier, EtFedDyn, is exploited which is an extension of FedDyn, with the key differences on using focal loss cost function to tackle severe class imbalances in the datasets, and on multi-stream network to exploit MRIs in different modalities. By combining EtFedDyn with domain mapping as the pre-processing and 3D scan-based post-processing, the proposed scheme makes 3D brain scan-based classification on datasets from different dataset owners. To examine whether the FL scheme could replace the central learning (CL) one, we then compare the classification performance between the proposed FL and the corresponding CL schemes. Furthermore, detailed empirical-based analysis were also conducted to exam the effect of using domain mapping, 3D scan-based post-processing, different cost functions and different FL schemes. Results: Experiments were done on two case studies: classification of glioma subtypes (IDH mutation and wild-type on TCGA and US datasets in case A) and glioma grades (high/low grade glioma HGG and LGG on MICCAI dataset in case B). The proposed FL scheme has obtained good performance on the test sets (85.46%, 75.56%) for IDH subtypes and (89.28%, 90.72%) for glioma LGG/HGG all averaged on five runs. Comparing with the corresponding CL scheme, the drop in test accuracy from the proposed FL scheme is small (−1.17%, −0.83%), indicating its good potential to replace the CL scheme. Furthermore, the empirically tests have shown that an increased classification test accuracy by applying: domain mapping (0.4%, 1.85%) in case A; focal loss function (1.66%, 3.25%) in case A and (1.19%, 1.85%) in case B; 3D post-processing (2.11%, 2.23%) in case A and (1.81%, 2.39%) in case B and EtFedDyn over FedAvg classifier (1.05%, 1.55%) in case A and (1.23%, 1.81%) in case B with fast convergence, which all contributed to the improvement of overall performance in the proposed FL scheme. Conclusion: The proposed FL scheme is shown to be effective in predicting glioma and its subtypes by using MR images from test sets, with great potential of replacing the conventional CL approaches for training deep networks. This could help hospitals to maintain their data privacy, while using a federated trained classifier with nearly similar performance as that from a centrally trained one. Further detailed experiments have shown that different parts in the proposed 3D FL scheme, such as domain mapping (make datasets more uniform) and post-processing (scan-based classification), are essential.
|
|
| 6. |
- Ali, Muhaddisa Barat, 1986, et al.
(författare)
-
Domain Mapping and Deep Learning from Multiple MRI Clinical Datasets for Prediction of Molecular Subtypes in Low Grade Gliomas
- 2020
-
Ingår i: Brain Sciences. - : MDPI AG. - 2076-3425. ; 10:7, s. 1-20
-
Tidskriftsartikel (refereegranskat)abstract
- Brain tumors, such as low grade gliomas (LGG), are molecularly classified which require the surgical collection of tissue samples. The pre-surgical or non-operative identification of LGG molecular type could improve patient counseling and treatment decisions. However, radiographic approaches to LGG molecular classification are currently lacking, as clinicians are unable to reliably predict LGG molecular type using magnetic resonance imaging (MRI) studies. Machine learning approaches may improve the prediction of LGG molecular classification through MRI, however, the development of these techniques requires large annotated data sets. Merging clinical data from different hospitals to increase case numbers is needed, but the use of different scanners and settings can affect the results and simply combining them into a large dataset often have a significant negative impact on performance. This calls for efficient domain adaption methods. Despite some previous studies on domain adaptations, mapping MR images from different datasets to a common domain without affecting subtitle molecular-biomarker information has not been reported yet. In this paper, we propose an effective domain adaptation method based on Cycle Generative Adversarial Network (CycleGAN). The dataset is further enlarged by augmenting more MRIs using another GAN approach. Further, to tackle the issue of brain tumor segmentation that requires time and anatomical expertise to put exact boundary around the tumor, we have used a tight bounding box as a strategy. Finally, an efficient deep feature learning method, multi-stream convolutional autoencoder (CAE) and feature fusion, is proposed for the prediction of molecular subtypes (1p/19q-codeletion and IDH mutation). The experiments were conducted on a total of 161 patients consisting of FLAIR and T1 weighted with contrast enhanced (T1ce) MRIs from two different institutions in the USA and France. The proposed scheme is shown to achieve the test accuracy of 74.81% on 1p/19q codeletion and 81.19% on IDH mutation, with marked improvement over the results obtained without domain mapping. This approach is also shown to have comparable performance to several state-of-the-art methods.
|
|
| 7. |
- Ali, Muhaddisa Barat, 1986, et al.
(författare)
-
Multi-stream Convolutional Autoencoder and 2D Generative Adversarial Network for Glioma Classification
- 2019
-
Ingår i: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). - Cham : Springer International Publishing. - 1611-3349 .- 0302-9743. ; 11678 LNCS, s. 234-245
-
Konferensbidrag (refereegranskat)abstract
- Diagnosis and timely treatment play an important role in preventing brain tumor growth. Deep learning methods have gained much attention lately. Obtaining a large amount of annotated medical data remains a challenging issue. Furthermore, high dimensional features of brain images could lead to over-fitting. In this paper, we address the above issues. Firstly, we propose an architecture for Generative Adversarial Networks to generate good quality synthetic 2D MRIs from multi-modality MRIs (T1 contrast-enhanced, T2, FLAIR). Secondly, we propose a deep learning scheme based on 3-streams of Convolutional Autoencoders (CAEs) followed by sensor information fusion. The rational behind using CAEs is that it may improve glioma classification performance (as comparing with conventional CNNs), since CAEs offer noise robustness and also efficient feature reduction hence possibly reduce the over-fitting. A two-round training strategy is also applied by pre-training on GAN augmented synthetic MRIs followed by refined-training on original MRIs. Experiments on BraTS 2017 dataset have demonstrated the effectiveness of the proposed scheme (test accuracy 92.04%). Comparison with several exiting schemes has provided further support to the proposed scheme.
|
|
| 8. |
- Ali, Muhaddisa Barat, 1986, et al.
(författare)
-
Prediction of glioma‑subtypes: comparison of performance on a DL classifier using bounding box areas versus annotated tumors
- 2022
-
Ingår i: BioMedical Engineering Online. - : Springer Science and Business Media LLC. - 1475-925X .- 2524-4426. ; 4
-
Tidskriftsartikel (refereegranskat)abstract
- Background: For brain tumors, identifying the molecular subtypes from magnetic resonance imaging (MRI) isdesirable, but remains a challenging task. Recent machine learning and deep learning (DL) approaches may help theclassification/prediction of tumor subtypes through MRIs. However, most of these methods require annotated datawith ground truth (GT) tumor areas manually drawn by medical experts. The manual annotation is a time consumingprocess with high demand on medical personnel. As an alternative automatic segmentation is often used. However, itdoes not guarantee the quality and could lead to improper or failed segmented boundaries due to differences in MRIacquisition parameters across imaging centers, as segmentation is an ill‑defined problem. Analogous to visual objecttracking and classification, this paper shifts the paradigm by training a classifier using tumor bounding box areas inMR images. The aim of our study is to see whether it is possible to replace GT tumor areas by tumor bounding boxareas (e.g. ellipse shaped boxes) for classification without a significant drop in performance.Method: In patients with diffuse gliomas, training a deep learning classifier for subtype prediction by employ‑ing tumor regions of interest (ROIs) using ellipse bounding box versus manual annotated data. Experiments wereconducted on two datasets (US and TCGA) consisting of multi‑modality MRI scans where the US dataset containedpatients with diffuse low‑grade gliomas (dLGG) exclusively.Results: Prediction rates were obtained on 2 test datasets: 69.86% for 1p/19q codeletion status on US dataset and79.50% for IDH mutation/wild‑type on TCGA dataset. Comparisons with that of using annotated GT tumor data fortraining showed an average of 3.0% degradation (2.92% for 1p/19q codeletion status and 3.23% for IDH genotype).Conclusion: Using tumor ROIs, i.e., ellipse bounding box tumor areas to replace annotated GT tumor areas for train‑ing a deep learning scheme, cause only a modest decline in performance in terms of subtype prediction. With moredata that can be made available, this may be a reasonable trade‑off where decline in performance may be counter‑acted with more data.
|
|
| 9. |
- Alipoor, Mohammad, 1983, et al.
(författare)
-
A Novel Framework for repeated measurements in diffusion tensor imaging
- 2016
-
Ingår i: 3rd (ACM) Int'l Conf. on Biomedical and Bioinformatics Engineering (ICBBE 2016). - New York, NY, USA : ACM. - 9781450348249 ; Part F125793, s. 1-6
-
Konferensbidrag (refereegranskat)abstract
- In the context of diffusion tensor imaging (DTI), the utility of making repeated measurements in each diffusion sensitizing direction has been the subject of numerous stud-ies. One can estimate the true signal value using either the raw complex-valued data or the real-valued magnitudesignal. While conventional methods focus on the former strategy, this paper proposes a new framework for acquiring/processing repeated measurements based on the latter strategy. The aim is to enhance the DTI processing pipeline by adding a diffusion signal estimator (DSE). This permits us to exploit the knowledge of the noise distribution to estimate the true signal value in each direction. An extensive study of the proposed framework, including theoretical analysis, experiments with synthetic data, performance evaluation and comparisons is presented.Our results show that the precision of estimated diffusionparameters is dependent on the number of available samplesand the manner in which the DSE accounts for noise. Theproposed framework improves the precision in estimationof diffusion parameters given a sufficient number of uniquemeasurements. This encourages future work with rich realdatasets and downstream applications.
|
|
| 10. |
- Alipoor, Mohammad, 1983, et al.
(författare)
-
Determinant of the information matrix: a new rotation invariant optimality metric to design gradient encoding schemes
- 2015
-
Ingår i: 12th IEEE International Symposium on Biomedical Imaging, ISBI 2015, Brooklyn, United States, 16-19 April 2015. - 1945-8452. - 9781479923748 ; , s. 462-465
-
Konferensbidrag (refereegranskat)abstract
- Minimum condition number (CN) gradient encoding schemewas introduced to diffusion MRI community more than adecade ago. It’s computation requires tedious numerical optimization which usually leads to sub-optimal solutions. TheCN does not reflect any benefits in acquiring more measurements, i.e. it’s optimal value is constant for any numberof measurements. Further, it is variable under rotation. Inthis paper we (i) propose an accurate method to computeminimum condition number scheme; and (ii) introduce determinant of the information matrix (DIM) as a new optimality metric that scales with number of measurements anddoes reflect what one would gain from acquiring more measurements. Theoretical analysis shows that DIM is rotationinvariant. Evaluations on state-of-the-art encoding schemesproves the relevance and superiority of the proposed metriccompared to condition number.
|
|
| 11. |
- Alipoor, Mohammad, 1983, et al.
(författare)
-
Fourth order tensor-based diffusion MRI signal modeling
- 2015
-
Ingår i: International symposium on biomedical imaging, White Matter Modeling Challenge. 16-19 April 2015, New York, USA..
-
Konferensbidrag (övrigt vetenskapligt/konstnärligt)abstract
- This abstract describes forth order tensor-based diffusion signal modeling as proposed in [1].
|
|
| 12. |
- Alipoor, Mohammad, 1983, et al.
(författare)
-
Icosahedral gradient encoding scheme for an arbitrary number of measurements
- 2015
-
Ingår i: International symposium on biomedical imaging. - 1945-8452. - 9781479923748 ; 2015-July, s. 959-962
-
Konferensbidrag (refereegranskat)abstract
- The icosahedral gradient encoding scheme (GES) is widelyused in diffusion MRI community due to its uniformly distributed orientations and rotationally invariant condition number. The major drawback with this scheme is that it is notavailable for arbitrary number of measurements. In this paper(i) we propose an algorithm to find the icosahedral schemefor any number of measurements. Performance of the obtained GES is evaluated and compared with that of Jones andtraditional icosahedral schemes in terms of condition number,standard deviation of the estimated fractional anisotropy anddistribution of diffusion sensitizing directions; and (ii) we introduce minimum eigenvalue of the information matrix as anew optimality metric to replace condition number. Unlikecondition number, it is proportional to the number of measurements and thus in agreement with the intuition that moremeasurements leads to more robust tensor estimation. Furthermore, it may independently be maximized to design GESsfor different diffusion imaging techniques.
|
|
| 13. |
- Alipoor, Mohammad, 1983, et al.
(författare)
-
K-Optimal Gradient Encoding Scheme for Fourth-Order Tensor-Based Diffusion Profile Imaging
- 2015
-
Ingår i: Biomed Research International. - : Hindawi Limited. - 2314-6133 .- 2314-6141.
-
Tidskriftsartikel (refereegranskat)abstract
- The design of an optimal gradient encoding scheme (GES) is a fundamental problem in diffusion MRI. It is well studied for the case of second-order tensor imaging (Gaussian diffusion). However, it has not been investigated for the wide range of non-Gaussian diffusion models. The optimal GES is the one that minimizes the variance of the estimated parameters. Such a GES can be realized by minimizing the condition number of the design matrix (K-optimal design). In this paper, we propose a new approach to solve the K-optimal GES design problem for fourth-order tensor-based diffusion profile imaging. The problem is a nonconvex experiment design problem. Using convex relaxation, we reformulate it as a tractable semidefinite programming problem. Solving this problem leads to several theoretical properties of K-optimal design: (i) the odd moments of the K-optimal design must be zero; (ii) the even moments of the K-optimal design are proportional to the total number of measurements; (iii) the K-optimal design is not unique, in general; and (iv) the proposed method can be used to compute the K-optimal design for an arbitrary number of measurements. Our Monte Carlo simulations support the theoretical results and show that, in comparison with existing designs, the K-optimal design leads to the minimum signal deviation.
|
|
| 14. |
- Alipoor, Mohammad, 1983, et al.
(författare)
-
On High Order Tensor-based Diffusivity Profile Estimation
- 2013
-
Ingår i: Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS. - 1557-170X. - 9781457702167 ; , s. 93-96, s. 4-
-
Konferensbidrag (refereegranskat)abstract
- Diffusion weighted magnetic resonance imaging (dMRI) is used to measure, in vivo, the self-diffusion of water molecules in biological tissues. High order tensors (HOTs) are used to model the apparent diffusion coefficient (ADC) profile at each voxel from the dMRI data. In this paper we propose: (i) A new method for estimating HOTs from dMRI data based on weighted least squares (WLS) optimization; and (ii) A new expression for computing the fractional anisotropy from a HOT that does not suffer from singularities and spurious zeros. We also present an empirical evaluation of the proposed method relative to the two existing methods based on both synthetic and real human brain dMRI data. The results show that the proposed method yields more accurate estimation than the competing methods.
|
|
| 15. |
- Alipoor, Mohammad, 1983, et al.
(författare)
-
Optimal Diffusion Tensor Imaging with Repeated Measurements
- 2013
-
Ingår i: Lecture Notes in Computer Science: Medical Image Computing and Computer-Assisted Intervention – MICCAI 2013. 16th International Conference, Nagoya, Japan, September 22-26, 2013, Proceedings, Part I. - Berlin, Heidelberg : Springer Berlin Heidelberg. - 0302-9743 .- 1611-3349. - 9783642408106 ; 8149, s. 687-694
-
Konferensbidrag (refereegranskat)abstract
- Several data acquisition schemes for diffusion MRI have been proposed and explored to date for the reconstruction of the 2nd order tensor. Our main contributions in this paper are: (i) the definition of a new class of sampling schemes based on repeated measurements in every sampling point; (ii) two novel schemes belonging to this class; and (iii) a new reconstruction framework for the second scheme. We also present an evaluation, based on Monte Carlo computer simulations, of the performances of these schemes relative to known optimal sampling schemes for both 2nd and 4th order tensors. The results demonstrate that tensor estimation by the proposed sampling schemes and estimation framework is more accurate and robust.
|
|
| 16. |
- Alipoor, Mohammad, 1983, et al.
(författare)
-
Optimal Experiment Design for Mono-Exponential Model Fitting: Application to Apparent Diffusion Coefficient Imaging
- 2015
-
Ingår i: BioMed Research International. - : Hindawi Limited. - 2314-6133 .- 2314-6141. ; 2015
-
Tidskriftsartikel (refereegranskat)abstract
- The mono-exponential model is widely used in quantitative biomedical imaging. Notable applications include apparent diffusion coefficient (ADC) imaging and pharmacokinetics.The application of ADC imaging to the detection of malignant tissue has in turn prompted several studies concerning optimal experiment design for mono-exponential model fitting. In this paper, we propose a new experiment design method that is based on minimizing the determinant of the covariance matrix of the estimated parameters (?-optimal design). In contrast to previous methods, ?-optimal design is independent of the imaged quantities. Applying this method to ADC imaging, we demonstrate its steady performance for the whole range of input variables (imaged parameters, number of measurements, range of ?-values). Using Monte Carlo simulations we show that the ?-optimal design outperforms existing experiment design methods in terms of accuracy and precision of the estimated parameters.
|
|
| 17. |
- Alipoor, Mohammad, 1983, et al.
(författare)
-
Optimal Gradient Encoding Schemes for Diffusion Tensor and Kurtosis Imaging
- 2016
-
Ingår i: IEEE transactions on Computational Imaging. - 2333-9403. ; 2:3, s. 375-391
-
Tidskriftsartikel (refereegranskat)abstract
- Diffusion-derived parameters find application in characterizing pathological and developmental changes in living tissues. Robust estimation of these parameters is important because they are used for medical diagnosis. An optimal gradient encoding scheme (GES) is one that minimizes the variance of the estimated diffusion parameters. This paper proposes a method for optimal GES design for two diffusion models: high-order diffusion tensor (HODT) imaging and diffusion kurtosis imaging (DKI). In both cases, the optimal GES design problem is formulated as a D-optimal (minimum determinant) experiment design problem. Then, using convex relaxation, it is reformulated as a semidefinite programming problem. Solving these problems we show that: 1) there exists a D-optimal solution for DKI that is simultaneously D-optimal for second- and fourth-order diffusion tensor imaging (DTI); 2) the traditionally used icosahedral scheme is approximately D-optimal for DTI and DKI; 3) the proposed D-optimal design is rotation invariant; 4) the proposed method can be used to compute the optimal design ($b$ -values and directions) for an arbitrary number of measurements and shells; and 5) using the proposed method one can obtain uniform distribution of gradient encoding directions for a typical number of measurements. Importantly, these theoretical findings provide the first mathematical proof of the optimality of uniformly distributed GESs for DKI and HODT imaging. The utility of the proposed method is further supported by the evaluation results and comparisons with with existing methods.
|
|
| 18. |
|
|
| 19. |
|
|
| 20. |
|
|
| 21. |
- Axelberg, Peter G.V. 1959, et al.
(författare)
-
AUTOMATIC CLASSIFICATION OF VOLTAGE EVENTS USING THE SUPPORT VECTOR MACHINE METHOD
- 2007
-
Ingår i: 19th International Conference on Electricity Distribution (SIRED 2007) , Vienna, Austria, 21-24 May, 2007.
-
Konferensbidrag (refereegranskat)abstract
- Statistically based classification systems need to be trained on a large number of training data in order to classify unseen data accurately. However, it is difficult to gather enough voltage events for the training purpose from real recordings. Therefore, a classification system trained to accurately classify real voltage events, but based on synthetic training data is highly in demand. This paper therefore proposes the design of a statistically based classification system trained on synthetic data. The paper gives also the results of conducted performance tests when the proposed classification system was trained to classify seven common types of voltage events. The experiments showed an overall detection rate of 81.6%, 91.9% and 99.5% respectively.
|
|
| 22. |
- Axelberg, Peter G.V., et al.
(författare)
-
Performance Tests of a Support Vector Machine used for Classification of Voltage Disturbances
- 2006
-
Ingår i: in proc. of 12th International conf. on Harmonics and Quality of Power (ICHQP 2006), Cascais, Portugal, Oct.1-5, 2006.
-
Konferensbidrag (refereegranskat)abstract
- This paper proposes a novel method for classifying voltage disturbances in electric power systems by using the Support Vector Machine (SVM) method. The proposed SVM classifier is designed to classify five common types of voltage disturbances and experiments have been conducted on recorded disturbances with good classification results. The proposed SVM classifier is also shown to be robust in terms of using training data and testing data that originate from two different power networks.
|
|
| 23. |
- Axelberg, Peter G.V. 1959, et al.
(författare)
-
Support Vector Machine for Classification of Voltage Disturbances
- 2007
-
Ingår i: accepted for publication in IEEE Transactions on Power Delivery. ; 22:3, s. 1297-1303, July, 2007
-
Tidskriftsartikel (refereegranskat)abstract
- The Support Vector Machine (SVM) is a powerful method for statistical classification of data used in a number of different applications. However, the usefulness of the method in a commercial available system is very much dependent on whether the SVM classifier can be pre-trained from a factory since it is not realistic that the SVM classifier must be trained by the customers themselves before it can be used. We first propose a novel SVM classification system for voltage disturbances. Our aim also includes investigating the performance of the proposed SVM classifier when the voltage disturbance data used for training and testing are originated from different sources. The data used in the experiments were originated from both real disturbances recorded in two different power networks and from synthetic data. The experimental results have shown excellent accuracy in classification when training data were originated from one power network and unseen testing data from another. High accuracy was also achieved when the SVM classifier was trained on data from a real power network and test data originated from synthetic data. Slightly less accuracy was achieved when the SVM classifier was trained on synthetic data and test data were originated from the power network.
|
|
| 24. |
- Axelberg, Peter G.V. 1959, et al.
(författare)
-
Trace of flicker sources by using the quantity of flicker power
- 2007
-
Ingår i: IEEE transactions on Power Delivery. ; 23:1, s. pp.465-471
-
Tidskriftsartikel (refereegranskat)abstract
- Industries that produce flicker are often placed close to each other and connected to the same power grid system. This implies that the measured flicker level at the point of common coupling (PCC) is a result of contribution from a number of different flicker sources. In a mitigation process it is essential to know which one of the flicker sources is the dominant one. We propose a method to determine the flicker propagations and trace the flicker sources by using flicker power measurements. Flicker power is considered as a quantity containing both sign and magnitude. The sign determines if a flicker source is placed downstream or upstream with respect to a given monitoring point and the magnitude is used to determine the propagation of flicker power throughout the power network and to trace the dominant flicker source. This paper covers the theoretical background of flicker power and describes a novel method for calculation of flicker power that can be implemented in a power network analyzer. Also conducted simulations and a field test based on the proposed method will be described in the paper.
|
|
| 25. |
|
|
| 26. |
|
|
| 27. |
|
|
| 28. |
|
|
| 29. |
- Backhouse, Andrew, 1978, et al.
(författare)
-
ML Nonlinear Smoothing for Image Segmentation and Its Relationship to The Mean Shift
- 2007
-
Ingår i: IEEE International conf. on Image Processing (ICIP '07).
-
Konferensbidrag (refereegranskat)abstract
- This paper addresses the issues of nonlinear edge-preserving image smoothing and segmentation. A ML-based approach is proposed which uses an iterative algorithm to solve the problem. First, assumptions about segments are made by describing the joint probability distribution of pixel positions and colours within segments. Based on these assumptions, an optimal smoothing algorithm is derived under the ML condition. By studying the derived algorithm, we show that the solution is related to a two-stage mean shift which is separated in space and range. This novel ML-based approach takes a new kernel function. Experiments have been conducted on a range of images to smooth and segment them. Visual results and evaluations with 2 objective criteria have shown that the proposed method has led to improved results which suffer from less over-segmentation than the standard mean-shift.
|
|
| 30. |
|
|
| 31. |
- Backhouse, Andrew, 1978, et al.
(författare)
-
Robust Object Tracking using Particle Filters and Multi-Region Mean Shift
- 2009
-
Ingår i: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). - Berlin, Heidelberg : Springer Berlin Heidelberg. - 1611-3349 .- 0302-9743. - 9783642104664 ; 5879, s. 11-403
-
Konferensbidrag (refereegranskat)abstract
- In this paper, we introduce a novel algorithm which buildsupon the combined anisotropic mean-shift and particle filter framework. The anisotropic mean-shift with 5 degrees of freedom, is extended to work on a partition of the object into concentric rings. This adds spatial information to the description of the object which makes the algorithm more resilient to occlusion and less susceptible to confusion with objects having similar color densities. Experiments conducted on videos containing deformable objects with long-term partial occlusion (or, short-term full occlusion) and intersection have shown robust tracking performance, especially in tracking objects with long term partial occlusion, short term full occlusion, close color background clutter, severe object deformation and fast changing motion. Comparisons with two existing methods have shown marked improvement in terms of robustness to occlusions, tightness and accuracy of tracked box, and tracking drifts.
|
|
| 32. |
- Bagheri, Azam, et al.
(författare)
-
A Framework Based on Machine Learning for Analytics of Voltage Quality Disturbances
- 2022
-
Ingår i: Energies. - : MDPI. - 1996-1073. ; 15:4
-
Tidskriftsartikel (refereegranskat)abstract
- This paper proposes a machine-learning-based framework for voltage quality analytics, where the space phasor model (SPM) of the three-phase voltages before, during, and after the event is applied as input data. The framework proceeds along with three main steps: (a) event extraction, (b) event characterization, and (c) additional information extraction. During the first step, it utilizes a Gaussian-based anomaly detection (GAD) technique to extract the event data from the recording. Principal component analysis (PCA) is adopted during the second step, where it is shown that the principal components correspond to the semi-minor and semi-major axis of the ellipse formed by the SPM. During the third step, these characteristics are interpreted to extract additional information about the underlying cause of the event. The performance of the framework was verified through experiments conducted on datasets containing synthetic and measured power quality events. The results show that the combination of semi-major axis, semi-minor axis, and direction of the major axis forms a sufficient base to characterize, classify, and eventually extract additional information from recorded event data.
|
|
| 33. |
- Bagheri, Azam, et al.
(författare)
-
A Robust Transform-Domain Deep Convolutional Network for Voltage Dip Classification
- 2018
-
Ingår i: IEEE Transactions on Power Delivery. - : Institute of Electrical and Electronics Engineers (IEEE). - 0885-8977 .- 1937-4208. ; 33:6, s. 2794-2802
-
Tidskriftsartikel (refereegranskat)abstract
- This paper proposes a novel method for voltage dip classification using deep convolutional neural networks. The main contributions of this paper include: (a) to propose a new effective deep convolutional neural network architecture for automatically learning voltage dip features, rather than extracting hand-crafted features; (b) to employ the deep learning in an effective two-dimensional transform domain, under space-phasor model (SPM), for efficient learning of dip features; (c) to characterize voltage dips by two-dimensional SPM-based deep learning, which leads to voltage dip features independent of the duration and sampling frequency of dip recordings; (d) to develop robust automatically-extracted features that are insensitive to training and test datasets measured from different countries/regions.Experiments were conducted on datasets containing about 6000 measured voltage dips spread over seven classes measured from several different countries. Results have shown good performance of the proposed method: average classification rate is about 97% and false alarm rate is about 0.50%. The test results from the proposed method are compared with the results from two existing dip classification methods. The proposed method is shown to out-perform these existing methods.
|
|
| 34. |
- Bagheri, Azam, et al.
(författare)
-
Big data from smart grids
- 2018
-
Ingår i: 2017 IEEE PES Innovative Smart Grid Technologies Conference Europe, ISGT-Europe 2017. - New York : Institute of Electrical and Electronics Engineers (IEEE). - 9781538619537
-
Konferensbidrag (refereegranskat)abstract
- This paper gives a general introduction to “Big Data” in general and to Big Data in smart grids in particular. Large amounts of data (Big Data) contains a lots of information, however developing the analytics to extract such information is a big challenge due to some of the particular characteristics of Big Data. This paper investigates some existing analytic algorithms, especially deep learning algorithms, as tools for handling Big Data. The paper also explains the potential for deep learning application in smart grids.
|
|
| 35. |
- Bagheri, Azam, 1982, et al.
(författare)
-
Estimation of frequency-dependent impedances in power grids by deep lstm autoencoder and random forest
- 2021
-
Ingår i: Energies. - : MDPI AG. - 1996-1073 .- 1996-1073. ; 14:13
-
Tidskriftsartikel (refereegranskat)abstract
- This paper proposes a deep-learning-based method for frequency-dependent grid impedance estimation. Through measurement of voltages and currents at a specific system bus, the estimate of the grid impedance was obtained by first extracting the sequences of the time-dependent features for the measured data using a long short-term memory autoencoder (LSTM-AE) followed by a random forest (RF) regression method to find the nonlinear map function between extracted features and the corresponding grid impedance for a wide range of frequencies. The method was trained via simulation by using time-series measurements (i.e., voltage and current) for different system parameters and verified through several case studies. The obtained results show that: (1) extracting the time-dependent features of the voltage/current data improves the performance of the RF regression method; (2) the RF regression method is robust and allows grid impedance estimation within 1.5 grid cycles; (3) the proposed method can effectively estimate the grid impedance both in steady state and in case of large transients like electrical faults.
|
|
| 36. |
- Bagheri, Azam, et al.
(författare)
-
Generative Adversarial Model-Guided Deep Active Learning for Voltage Dip Labelling
- 2019
-
Ingår i: 2019 IEEE Milan PowerTech. - : IEEE.
-
Konferensbidrag (refereegranskat)abstract
- In many real applications, the ground truths of class labels from voltage dip sequences used for training a voltage dip classification system are unknown, and require manual labelling by human experts. This paper proposes a novel deep active learning method for automatic labelling of voltage dip sequences used for the training process. We propose a novel deep active learning method, guided by a generative adversarial network (GAN), where the generator is formed by modelling data with a Gaussian mixture model and provides the estimated probability distribution function (pdf) where the query criterion of the deep active learning method is built upon. Furthermore, the discriminator is formed by a support vector machine (SVM). The proposed method has been tested on a voltage dip dataset (containing 916 dips) measured in a European country. The experiments have resulted in good performance (classification rate 83% and false alarm 3.2%), which have demonstrated the effectiveness of the proposed method.
|
|
| 37. |
- Bagheri, Azam, et al.
(författare)
-
Improved characterization of multi-stage voltage dips based on the space phasor model
- 2018
-
Ingår i: Electric power systems research. - : Elsevier. - 0378-7796 .- 1873-2046. ; 154, s. 319-328
-
Tidskriftsartikel (refereegranskat)abstract
- This paper proposes a method for characterizing voltage dips based on the space phasor model of the three phase-to-neutral voltages, instead of the individual voltages. This has several advantages. Using a K-means clustering algorithm, a multi-stage dip is separated into its individual event segments directly instead of first detecting the transition segments. The logistic regression algorithm fits the best single-segment characteristics to every individual segment, instead of extreme values being used for this, as in earlier methods. The method is validated by applying it to synthetic and measured dips. It can be generalized for application to both single- and multi-stage dips.
|
|
| 38. |
- Balouji, Ebrahim, 1985, et al.
(författare)
-
A LSTM-based Deep Learning Method with Application to Voltage Dip Classification
- 2018
-
Ingår i: 2018 18TH INTERNATIONAL CONFERENCE ON HARMONICS AND QUALITY OF POWER (ICHQP). - Piscataway, NJ : Institute of Electrical and Electronics Engineers (IEEE). - 2164-0610. - 9781538605172 - 9781538605172
-
Konferensbidrag (refereegranskat)abstract
- In this paper, a deep learning (DL)-based method for automatic feature extraction and classification of voltage dips is proposed. The method consists of a dedicated architecture of Long Short-Term Memory (LSTM), which is a special type of Recurrent Neural Networks (RNNs). A total of 5982 three-phase one-cycle voltage dip RMS sequences, measured from several countries, has been used in our experiments. Our results have shown that the proposedmethod is able to classify the voltage dips from learned features in LSTM, with 93.40% classification accuracy on the test data set. The developed architecture is shown to be novel for feature learning and classification of voltage dips. Different from the conventional machine learning methods, the proposed method is able to learn dip features without requiring transition-event segmentation, selecting thresholds, and using expert rules or human expert knowledge, when a large amount of measurement data is available. This opens a new possibility of exploiting deep learning technology for power quality data analytics and classification.
|
|
| 39. |
- Bengtsson, Tomas, 1983, et al.
(författare)
-
Regularized Optimization for Joint Super-Resolution and High Dynamoc Range Image Reconstruction in a Percetually Uniform Domain.
- 2012
-
Ingår i: ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings. - 1520-6149. - 9781467300469 ; , s. 1097-1100
-
Konferensbidrag (refereegranskat)abstract
- This paper discusses resolution enhancement of a set of images with varying exposure durations, having a high combined dynamic range. So far, little has been said in relation to the Human Visual System when it comes to Super-Resolution and High Dynamic Range fusion, unlike the case for traditional Super-Resolution where errors are measured with respect to human perception in the pixel domain. We propose a Super-Resolution method in the L*a*b* domain to bridge the gap and present some image reconstruction results.
|
|
| 40. |
- Bengtsson, Tomas, 1983, et al.
(författare)
-
Super-resolution reconstruction of high dynamic range images in a perceptually uniform domain
- 2013
-
Ingår i: Optical Engineering. - 1560-2303 .- 0091-3286. ; 52:10
-
Tidskriftsartikel (refereegranskat)abstract
- Super resolution is a signal processing method that utilizesinformation from multiple degraded images of the same scene in order to reconstruct an image with enhanced spatial resolution. The method is typically employed on similarly exposed pixel valued images, but it can be extended to differently exposed images with a high combined dynamicrange. We propose a novel formulation of the joint super-resolution, high dynamic range image reconstruction problem, using an image domain in which the residual function of the inverse problem relates to the perception of the human visual system. Simulated results are presented,including a comparison with a conventional method, demonstrating that the proposed approach avoids some severe reconstruction artifacts.
|
|
| 41. |
- Bengtsson, Tomas, 1983, et al.
(författare)
-
Super-resolution reconstruction of high dynamic range images with perceptual weighting of errors
- 2013
-
Ingår i: ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings. - 1520-6149. - 9781479903566 ; , s. 2212-2216
-
Konferensbidrag (refereegranskat)abstract
- Super-Resolution and High Dynamic Range image reconstruction are two different signal processing techniques that share in common that they utilize information from multiple observations of the same scene to enhance visual image quality. In this paper, both techniques are merged in a common model, and the focus is to solve the reconstruction problem in a suitable image domain, which relates to the perception of the Human Visual System. Simulated results are presented, including a comparison with a conventional method, demonstrating the benefits of the proposed approach, in this case avoiding some severe reconstruction artifacts.
|
|
| 42. |
- Berlijn, Sonja M., et al.
(författare)
-
Laboratory Tests and Web Based Surveillance to Determine the Ice- and Snow Performance of Insulators
- 2007
-
Ingår i: IEEE Transactions on Dielectrics and Electrical Insulation, Special Issue on Flashover of Ice or Snow-Covered Insulators. ; 14:6, s. 1373-1380, 2007
-
Tidskriftsartikel (refereegranskat)abstract
- To be able to determine, verify and monitor the ice- and snow performance of different insulation solutions, laboratory test methods and an on-site, on-line web based surveillance system are needed. The method for determining the ice performance in laboratory conditions, Ice Progressive Stress (IPS) method, is described in this paper. Further it is described how to use the results of this type of test to estimate statistically the performance of complete overhead line equipped by different insulators. To verify the laboratory ice- and snow test method, to get an idea about the type and number of ice and snow events actually occurring in service and to get more information about ice and snow phenomena in real life an on-site on-line web based surveillance system was designed and built. This sophisticated system, including the automatic image analysis and used statistical tools is described in this paper. Besides the description of the laboratory test method and the surveillance system, service experience, pictures and interesting results obtained so far are also presented.
|
|
| 43. |
- Berlijn, Sonja, et al.
(författare)
-
Practical Applications of Automatic Image Analysis of Overhead Power Lines
- 2014
-
Ingår i: INMR (Independent Transmission and Distribution Information Resource). ; :103, Q1, s. 72-75
-
Tidskriftsartikel (refereegranskat)abstract
- Norwegian transmission system operator, Statnett, has a network comprising more than 10,000 km of overhead lines operating at from 132 kV to 420 kV. For several years now, this TSO has conducted research at a special WAP test facility located near Oslo, where different insulator options, a weather station and special monitoring equipment have been installed, including web cameras and ice load monitors. Recently, the test station has also been used to collect information about comparative 'visibility' of different insulators since this has become an increasingly important parameter in overhead line design in Norway and indeed across Europe.This article, conducted by Sonja Berlijn of Statnett, Igor Gutman of STRI and Irene Gu of the Chalmers University of Technology in Gothenburg, presents novel applications of automatic analysis used in high voltage engineering. These include detecting and measuring snow/ice coverage on insulators, estimating insulator swing angles and computing objective values of how different insulator strings stand out against specific backgrounds.
|
|
| 44. |
|
|
| 45. |
|
|
| 46. |
|
|
| 47. |
|
|
| 48. |
- Bollen, Math, et al.
(författare)
-
Classification of Underlying Causes of Power Quality Disturbances: Deterministic versus Statistical Methods
- 2007
-
Ingår i: Eurasip Journal on Applied Signal Processing. - : Springer Science and Business Media LLC. - 1110-8657 .- 1687-0433 .- 1687-6172 .- 1687-6180. ; , s. 17 pages (Article ID 79747)-
-
Tidskriftsartikel (refereegranskat)abstract
- This paper presents the two main types of classification methods for power quality disturbances based on underlying causes: deterministic classification, giving an expert system as an example, and statistical classification, with support vector machines as an example. An expert system is suitable when one has limited amount of data and sufficient power system expert knowledge, however its application requires a set of threshold values. Statistical methods are suitable when large amount of data is available for training. Two important issues to guarantee the effectiveness of a classifier, data segmentation and featureextraction, are discussed. Segmentation of a sequence of data recording is pre-processing to partition the datainto segments each representing a duration containing either an event or transition between two events. Extraction of features is applied to each segment individually. Some useful features and their effectiveness are then discussed. Some experimental results are included for demonstrating theeffectiveness of both systems. Finally, conclusions are given together with the discussion of some future research directions.
|
|
| 49. |
|
|
| 50. |
|
|
