Publications
Results
- Showing results for:
- Reset all filters
Search results
-
Conference paperWang C, Papanastasiou G, Tsaftaris S, et al., 2019,
TPSDicyc: Improved deformation invariant cross-domain medical image synthesis
, Second International Workshop, MLMIR 2019, Held in Conjunction with MICCAI 2019, Publisher: Springer International Publishing, Pages: 245-254, ISSN: 0302-9743Cycle-consistent generative adversarial network (CycleGAN) has been widely used for cross-domain medical image systhesis tasks particularly due to its ability to deal with unpaired data. However, most CycleGAN-based synthesis methods can not achieve good alignment between the synthesized images and data from the source domain, even with additional image alignment losses. This is because the CycleGAN generator network can encode the relative deformations and noises associated to different domains. This can be detrimental for the downstream applications that rely on the synthesized images, such as generating pseudo-CT for PET-MR attenuation correction. In this paper, we present a deformation invariant model based on the deformation-invariant CycleGAN (DicycleGAN) architecture and the spatial transformation network (STN) using thin-plate-spline (TPS). The proposed method can be trained with unpaired and unaligned data, and generate synthesised images aligned with the source data. Robustness to the presence of relative deformations between data from the source and target domain has been evaluated through experiments on multi-sequence brain MR data and multi-modality abdominal CT and MR data. Experiment results demonstrated that our method can achieve better alignment between the source and target data while maintaining superior image quality of signal compared to several state-of-the-art CycleGAN-based methods.
-
Conference paperChen J, Zhang H, Zhang Y, et al., 2019,
Discriminative consistent domain generation for semi-supervised learning
, International Conference on Medical Image Computing and Computer-Assisted Intervention, Publisher: Springer International Publishing, Pages: 595-604, ISSN: 0302-9743Deep learning based task systems normally rely on a large amount of manually labeled training data, which is expensive to obtain and subject to operator variations. Moreover, it does not always hold that the manually labeled data and the unlabeled data are sitting in the same distribution. In this paper, we alleviate these problems by proposing a discriminative consistent domain generation (DCDG) approach to achieve a semi-supervised learning. The discriminative consistent domain is achieved by a double-sided domain adaptation. The double-sided domain adaptation aims to make a fusion of the feature spaces of labeled data and unlabeled data. In this way, we can fit the differences of various distributions between labeled data and unlabeled data. In order to keep the discriminativeness of generated consistent domain for the task learning, we apply an indirect learning for the double-sided domain adaptation. Based on the generated discriminative consistent domain, we can use the unlabeled data to learn the task model along with the labeled data via a consistent image generation. We demonstrate the performance of our proposed DCDG on the late gadolinium enhancement cardiac MRI (LGE-CMRI) images acquired from patients with atrial fibrillation in two clinical centers for the segmentation of the left atrium anatomy (LA) and proximal pulmonary veins (PVs). The experiments show that our semi-supervised approach achieves compelling segmentation results, which can prove the robustness of DCDG for the semi-supervised learning using the unlabeled data along with labeled data acquired from a single center or multicenter studies.
-
Journal articlePoo SX, Tham CSW, Smith C, et al., 2019,
IgG4-related disease in a multi-ethnic community: Clinical characteristics and association with malignancy
, QJM: An International Journal of Medicine, Vol: 112, Pages: 763-769, ISSN: 1460-2393BackgroundImmunoglobulin-G4-related disease (IgG4-RD) is a recently recognised fibro-inflammatory condition that can affect multiple organs. Despite growing interest in this condition, the natural history and management of IgG4-RD remain poorly understood.AimTo describe the clinical characteristics, treatment and outcomes of IgG4-RD in a multi-ethnic UK cohort, and investigate its possible association with malignancy.DesignRetrospective analysis of case-note and electronic data.MethodsCases were identified from sub-specialty cohorts and a systematic search of an NHS trust histopathology database using ‘IgG4’ or ‘inflammatory pseudotumour’ as search terms. Electronic records, imaging and histopathology reports were reviewed.Results66 identified cases of IgG4-RD showed a similar multi-ethnic spread to the local population of North West London. The median age was 59 years and 71% of patients were male. Presenting symptoms relating to mass effect of a lesion were present in 48% of cases and the mean number of organs involved was 2.4. 10 patients had reported malignancies with 6 of these being haematological. 83% of those treated with steroids had good initial response, however 50% had relapsing-remitting disease. Rituximab was administered in 11 cases and all achieved an initial serological response. Despite this, 7 patients subsequently relapsed after a mean duration of 11 months and 4 progressed despite treatment.ConclusionsWe report a large UK-based cohort of IgG4-RD that shows no clear ethnic predisposition and a wide range of affected organs. We discuss the use of serum IgG4 concentrations as a disease marker in IgG4-RD, the association with malignant disease and outcomes according to differing treatment regimens.
-
Conference paperAli AR, Li J, O'Shea SJ, et al., 2019,
A Deep Learning Based Approach to Skin Lesion Border Extraction with a Novel Edge Detector in Dermoscopy Images
Lesion border detection is considered a crucial step in diagnosing skin cancer. However, performing such a task automatically is challenging due to the low contrast between the surrounding skin and lesion, ambiguous lesion borders, and the presence of artifacts such as hair. In this paper we propose a two-stage approach for skin lesion border detection: (i) segmenting the skin lesion dermoscopy image using U-Net, and (ii) extracting the edges from the segmented image using a novel approach we call FuzzEdge. The proposed approach is compared with another published skin lesion border detection approach, and the results show that our approach performs better in detecting the main borders of the lesion and is more robust to artifacts that might be present in the image. The approach is also compared with the manual border drawings of a dermatologist, resulting in an average Dice similarity of 87.7%.
-
Journal articleZhang N, Yang G, Gao Z, et al., 2019,
Deep learning for diagnosis of chronic myocardial infarction on nonenhanced cardiac cine MRI
, Radiology, Vol: 294, Pages: 52-60, ISSN: 0033-8419BackgroundRenal impairment is common in patients with coronary artery disease and, if severe, late gadolinium enhancement (LGE) imaging for myocardial infarction (MI) evaluation cannot be performed.PurposeTo develop a fully automatic framework for chronic MI delineation via deep learning on non–contrast material–enhanced cardiac cine MRI.Materials and MethodsIn this retrospective single-center study, a deep learning model was developed to extract motion features from the left ventricle and delineate MI regions on nonenhanced cardiac cine MRI collected between October 2015 and March 2017. Patients with chronic MI, as well as healthy control patients, had both nonenhanced cardiac cine (25 phases per cardiac cycle) and LGE MRI examinations. Eighty percent of MRI examinations were used for the training data set and 20% for the independent testing data set. Chronic MI regions on LGE MRI were defined as ground truth. Diagnostic performance was assessed by analysis of the area under the receiver operating characteristic curve (AUC). MI area and MI area percentage from nonenhanced cardiac cine and LGE MRI were compared by using the Pearson correlation, paired t test, and Bland-Altman analysis.ResultsStudy participants included 212 patients with chronic MI (men, 171; age, 57.2 years ± 12.5) and 87 healthy control patients (men, 42; age, 43.3 years ± 15.5). Using the full cardiac cine MRI, the per-segment sensitivity and specificity for detecting chronic MI in the independent test set was 89.8% and 99.1%, respectively, with an AUC of 0.94. There were no differences between nonenhanced cardiac cine and LGE MRI analyses in number of MI segments (114 vs 127, respectively; P = .38), per-patient MI area (6.2 cm2 ± 2.8 vs 5.5 cm2 ± 2.3, respectively; P = .27; correlation coefficient, r = 0.88), and MI area percentage (21.5% ± 17.3 vs 18.5% ± 15.4; P = .17; correlation coefficient, r = 0.89).ConclusionThe proposed deep learning f
-
Journal articleKalna V, Yang Y, Peghaire C, et al., 2019,
The transcription factor ERG regulates super-enhancers associated with an endothelial-specific gene expression program
, Circulation Research, Vol: 124, Pages: 1337-1349, ISSN: 0009-7330Rationale:The ETS (E-26 transformation-specific) transcription factor ERG (ETS-related gene) is essential for endothelial homeostasis, driving expression of lineage genes and repressing proinflammatory genes. Loss of ERG expression is associated with diseases including atherosclerosis. ERG’s homeostatic function is lineage-specific, because aberrant ERG expression in cancer is oncogenic. The molecular basis for ERG lineage-specific activity is unknown. Transcriptional regulation of lineage specificity is linked to enhancer clusters (super-enhancers).Objective:To investigate whether ERG regulates endothelial-specific gene expression via super-enhancers.Methods and Results:Chromatin immunoprecipitation with high-throughput sequencing in human umbilical vein endothelial cells showed that ERG binds 93% of super-enhancers ranked according to H3K27ac, a mark of active chromatin. These were associated with endothelial genes such as DLL4 (Delta-like protein 4), CLDN5 (claudin-5), VWF (von Willebrand factor), and CDH5 (VE-cadherin). Comparison between human umbilical vein endothelial cell and prostate cancer TMPRSS2 (transmembrane protease, serine-2):ERG fusion-positive human prostate epithelial cancer cell line (VCaP) cells revealed distinctive lineage-specific transcriptome and super-enhancer profiles. At a subset of endothelial super-enhancers (including DLL4 and CLDN5), loss of ERG results in significant reduction in gene expression which correlates with decreased enrichment of H3K27ac and MED (Mediator complex subunit)-1, and reduced recruitment of acetyltransferase p300. At these super-enhancers, co-occupancy of GATA2 (GATA-binding protein 2) and AP-1 (activator protein 1) is significantly lower compared with super-enhancers that remained constant following ERG inhibition. These data suggest distinct mechanisms of super-enhancer regulation in endothelial cells and highlight the unique role of ERG in controlling a core subset of super-enhancers. Most disease-assoc
-
Journal articleZhang L, Yang G, Ye X, 2019,
Automatic skin lesion segmentation by coupling deep fully convolutional networks and shallow network with textons
, Journal of Medical Imaging, Vol: 6, ISSN: 2329-4302Segmentation of skin lesions is an important step in computer-aided diagnosis of melanoma; it is also a very challenging task due to fuzzy lesion boundaries and heterogeneous lesion textures. We present a fully automatic method for skin lesion segmentation based on deep fully convolutional networks (FCNs). We investigate a shallow encoding network to model clinically valuable prior knowledge, in which spatial filters simulating simple cell receptive fields function in the primary visual cortex (V1) is considered. An effective fusing strategy using skip connections and convolution operators is then leveraged to couple prior knowledge encoded via shallow network with hierarchical data-driven features learned from the FCNs for detailed segmentation of the skin lesions. To our best knowledge, this is the first time the domain-specific hand craft features have been built into a deep network trained in an end-to-end manner for skin lesion segmentation. The method has been evaluated on both ISBI 2016 and ISBI 2017 skin lesion challenge datasets. We provide comparative evidence to demonstrate that our newly designed network can gain accuracy for lesion segmentation by coupling the prior knowledge encoded by the shallow network with the deep FCNs. Our method is robust without the need for data augmentation or comprehensive parameter tuning, and the experimental results show great promise of the method with effective model generalization compared to other state-of-the-art-methods.
-
Conference paperWang C, MacGillivray T, Macnaught G, et al., 2019,
A Two-Stage U-Net Model for 3D Multi-class Segmentation on Full-Resolution Cardiac Data
, Pages: 191-199, ISSN: 0302-9743Deep convolutional neural networks (CNNs) have achieved state-of-the-art performances for multi-class segmentation of medical images. However, a common problem when dealing with large, high resolution 3D data is that the volumes input into the deep CNNs has to be either cropped or downsampled due to limited memory capacity of computing devices. These operations can lead to loss of resolution and class imbalance in the input data batches, thus downgrade the performances of segmentation algorithms. Inspired by the architecture of image super-resolution CNN (SRCNN), we propose a two-stage modified U-Net framework that simultaneously learns to detect a ROI within the full volume and to classify voxels without losing the original resolution. Experiments on a variety of multi-modal 3D cardiac images have demonstrated that this framework shows better segmentation performances than state-of-the-art Deep CNNs with trained with the same similarity metrics.
-
Conference paperLi M, Dong S, Zhang K, et al., 2019,
Deep Learning intra-image and inter-images features for Co-saliency detection
In this paper, we propose a novel deep end-to-end co-saliency detection approach to extract common salient objects from images group. The existing approaches rely heavily on manually designed metrics to characterize co-saliency. However, these methods are so subjective and not flexible enough that leads to poor generalization ability. Furthermore, most approaches separate the process of single image features and group images features extraction, which ignore the correlation between these two features that can promote the model performance. The proposed approach solves these two problems by multistage representation to extract features based on high-spatial resolution CNN. In addition, we utilize the modified CAE to explore the learnable consistency. Finally, the intra-image contrast and the inter-images consistency are fused to generate the final co-saliency maps automatically among group images by multistage learning. Experiment results demonstrate the effectiveness and superiority of our approach beyond the state-of-the-art methods.
-
Conference paperDong S, Gao Z, Sun S, et al., 2019,
Holistic and deep feature pyramids for saliency detection
Saliency detection has been increasingly gaining research interest in recent years since many computer vision applications need to derive object attentions from images in the first steps. Multi-scale awareness of the saliency detector becomes essential to find thin and small attention regions as well as keeping high-level semantics. In this paper, we propose a novel holistic and deep feature pyramid neural network architecture that can leverage multi-scale semantics in feature encoding stage and saliency region prediction (decoding) stage. In the encoding stage, we exploit multi-scale and pyramidal hierarchy of feature maps via the densely connected network with variable-size dilated convolutions as well as a pyramid pooling. In the decoding stage, we fuse multi-level feature maps via up-sampling and convolution. In addition, we utilize the multi-level deep supervision via plugging in loss functions at every feature fusion level. Multi-loss supervision regularizes weights searching space among different tasks minimizing over-fitting and enhances gradient signal during backpropagation, and thus enables us training the network from scratch. This architecture builds an inherent multi-level semantic pyramidal feature maps at different scales and enhances model's capability in the saliency detection task. We validated our approach on six benchmark datasets and compared with eleven state-of-the-art methods. The results demonstrated that the design effectiveness and our approach outperformed the compared methods.
This data is extracted from the Web of Science and reproduced under a licence from Thomson Reuters. You may not copy or re-distribute this data in whole or in part without the written consent of the Science business of Thomson Reuters.