| 1. |
- Ge, Chenjie, 1991, et al.
(författare)
-
Enlarged Training Dataset by Pairwise GANs for Molecular-Based Brain Tumor Classification
- 2020
-
Ingår i: IEEE Access. - 2169-3536 .- 2169-3536. ; 8:1, s. 22560-22570
-
Tidskriftsartikel (refereegranskat)abstract
- This paper addresses issues of brain tumor subtype classification using Magnetic Resonance Images (MRIs) from different scanner modalities like T1 weighted, T1 weighted with contrast-enhanced, T2 weighted and FLAIR images. Currently most available glioma datasets are relatively moderate in size, and often accompanied with incomplete MRIs in different modalities. To tackle the commonly encountered problems of insufficiently large brain tumor datasets and incomplete modality of image for deep learning, we propose to add augmented brain MR images to enlarge the training dataset by employing a pairwise Generative Adversarial Network (GAN) model. The pairwise GAN is able to generate synthetic MRIs across different modalities. To achieve the patient-level diagnostic result, we propose a post-processing strategy to combine the slice-level glioma subtype classification results by majority voting. A two-stage course-to-fine training strategy is proposed to learn the glioma feature using GAN-augmented MRIs followed by real MRIs. To evaluate the effectiveness of the proposed scheme, experiments have been conducted on a brain tumor dataset for classifying glioma molecular subtypes: isocitrate dehydrogenase 1 (IDH1) mutation and IDH1 wild-type. Our results on the dataset have shown good performance (with test accuracy 88.82%). Comparisons with several state-of-the-art methods are also included.
|
|
| 2. |
- 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.
|
|
| 3. |
|
|
| 4. |
- Petersson, Jesper, 1974
(författare)
-
Medicine At A Distance In Sweden: Spatiotemporal Matters In Accomplishing Working Telemedicine
- 2011
-
Ingår i: Science Studies. - 0786-3012. ; 24:2, s. 43-62
-
Tidskriftsartikel (refereegranskat)abstract
- This paper examines the accomplishment of making technology work, using the discourse around telemedicine in Swedish healthcare during 1994-2003. The paper will compare four projects launched in the mid-1990s and policymakers’ visions of healthcare through telemedicine. I will employ a sociotechnical approach developed within Actor-Network Theory that understands functioning technology not as something intrinsic but as an outcome of an ongoing process of negotiations. In the paper, I will extend the sociotechnical approach of what constitutes working technology to include spatiotemporal matters. I will also approach the closely related issue of space that has become a concern of Actor-Network Theory scholars interested in the accomplishment and continued workings of technology as it travels. In this discussion, an emphasis on fixed relations (network space) has been challenged by investigations into changing relations (fluid space). This paper suggests that in order to travel well, technology must be both fixed and fluid.⁰
|
|
| 5. |
|
|
| 6. |
- 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.
|
|
| 7. |
- Reza, Salim, 1985-, et al.
(författare)
-
Smart dosimetry by pattern recognition using a single photon counting detector system in time over threshold mode
- 2012
-
Ingår i: Journal of Instrumentation. - 1748-0221 .- 1748-0221. ; 7:1
-
Tidskriftsartikel (refereegranskat)abstract
- The function of a dosimeter is to determine the absorbed dose of radiation, for those cases in which, generally, the particular type of radiation is already known. Lately, a number of applications have emerged in which all kinds of radiation are absorbed and are sorted by pattern recognition, such as the Medipix2 application in [1]. This form of smart dosimetry enables measurements where not only the total dosage is measured, but also the contributions of different types of radiation impacting upon the detector surface. Furthermore, the use of a photon counting system, where the energy deposition can be measured in each individual pixel, ensures measurements with a high degree of accuracy in relation to the pattern recognition. In this article a Timepix [2] detector system has been used in the creation of a smart dosimeter for Alpha, Beta and Gamma radiation. When a radioactive particle hits the detector surface it generates charge clusters and those impacting upon the detector surface are read out and image processing algorithms are then used to classify each charge cluster. The individual clusters are calculated and as a result, the dosage for each type of radiation is given. In some cases, several particles can impact in roughly the same place, forming overlapping clusters. In order to handle this problem, a cluster separation method has been added to the pattern recognition algorithm. When the clusters have been separated, they are classified by shape and sorted into the correct type of radiation. The algorithms and methods used in this dosimeter have been developed so as to be simple and computationally effective, in order to enable implementation on a portable device. © 2012 IOP Publishing Ltd and SISSA.
|
|
| 8. |
- Wang, Chunliang, 1980-, et al.
(författare)
-
An interactive software module for visualizing coronary arteries in CT angiography
- 2008
-
Ingår i: International Journal of Computer Assisted Radiology and Surgery. - Heidelberg/Berlin : Springer. - 1861-6410 .- 1861-6429. ; 3:1-2, s. 11-18
-
Tidskriftsartikel (refereegranskat)abstract
- A new software module for coronary artery segmentation and visualization in CT angiography (CTA) datasets is presented, which aims to interactively segment coronary arteries and visualize them in 3D with maximum intensity projection (MIP) and volume rendering (VRT).Materials and Methods: The software was built as a plug-in for the open-source PACS workstation OsiriX. The main segmentation function is based an optimized “virtual contrast injection” algorithm, which uses fuzzy connectedness of the vessel lumen to separate the contrast-filled structures from each other. The software was evaluated in 42 clinical coronary CTA datasets acquired with 64-slice CT using isotropic voxels of 0.3–0.5 mm.Results: The median processing time was 6.4 min, and 100% of main branches (right coronary artery, left circumflex artery and left anterior descending artery) and 86.9% (219/252) of visible minor branches were intact. Visually correct centerlines were obtained automatically in 94.7% (321/339) of the intact branches.Conclusion: The new software is a promising tool for coronary CTA post-processing providing good overviews of the coronary artery with limited user interaction on low-end hardware, and the coronary CTA diagnosis procedure could potentially be more time-efficient than using thin-slab technique.
|
|
| 9. |
- Wang, Chunliang, 1980-, et al.
(författare)
-
Integrating automatic and interactive method for coronary artery segmentation : let PACS workstation think ahead
- 2010
-
Ingår i: International Journal of Computer Assisted Radiology and Surgery. - : Springer Science and Business Media LLC. - 1861-6410 .- 1861-6429. ; 5:3, s. 275-285
-
Tidskriftsartikel (refereegranskat)abstract
- Purpose: To provide an efficient method to extract useful information from the increasing amount of coronary CTA.Methods: A quantitative coronary CTA analysis tool was built on OsiriX, which integrates both fully automatic and interactive methods for coronary artery extraction. The computational power of an ordinary PC is exploited by running the non-supervised coronary artery segmentation and centerline tracking in the background as soon as the images are received. When the user opens the data, the software provides a real-time interactive analysis environment.Results: The average overlap between the centerline created in our software and the reference standard was 96.0%. The average distance between them was 0.38 mm. The automatic procedure runs for 3-5 min as a single-thread application in background. Interactive processing takes 3 min in average.Conclusion: In preliminary experiments, the software achieved higher efficiency than the former interactive method, and reasonable accuracy compared to manual vessel extraction.
|
|
| 10. |
- Ge, Chenjie, 1991, et al.
(författare)
-
Multi-Stream Multi-Scale Deep Convolutional Networks for Alzheimer's Disease Detection using MR Images
- 2019
-
Ingår i: Neurocomputing. - : Elsevier BV. - 0925-2312 .- 1872-8286. ; 350, s. 60-69
-
Tidskriftsartikel (refereegranskat)abstract
- This paper addresses the issue of Alzheimer's disease (AD) detection from Magnetic Resonance Images (MRIs). Existing AD detection methods rely on global feature learning from the whole brain scans, while depending on the tissue types, AD related features in dierent tissue regions, e.g. grey matter (GM), white matter (WM), and cerebrospinal uid (CSF), show different characteristics. In this paper, we propose a deep learning method for multi-scale feature learning based on segmented tissue areas. A novel deep 3D multi-scale convolutional network scheme is proposed to generate multi-resolution features for AD detection. The proposed scheme employs several parallel 3D multi-scale convolutional networks, each applying to individual tissue regions (GM, WM and CSF) followed by feature fusions. The proposed fusion is applied in two separate levels: the rst level fusion is applied on different scales within the same tissue region, and the second level is on dierent tissue regions. To further reduce the dimensions of features and mitigate overtting, a feature boosting and dimension reduction method, XGBoost, is utilized before the classication. The proposed deep learning scheme has been tested on a moderate open dataset of ADNI (1198 scans from 337 subjects), with excellent test performance on randomly partitioned datasets (best 99.67%, average 98.29%), and good test performance on subject-separated partitioned datasets (best 94.74%, average 89.51%). Comparisons with state-of-the-art methods are also included.
|
|
