Domain adaptation is crucial for transferring the knowledge from the source labeled CT dataset to the target unlabeled MR dataset in abdominal multi-organ segmentation. Meanwhile, it is highly ...desirable to avoid the high annotation cost related to the target dataset and protect the source dataset privacy. Therefore, we propose an effective source-free unsupervised domain adaptation method for cross-modality abdominal multi-organ segmentation without source dataset access. The proposed framework comprises two stages. In the first stage, the feature map statistics-guided model adaptation combined with entropy minimization is developed to help the top segmentation network reliably segment the target images. The pseudo-labels output from the top segmentation network are used to guide the style compensation network to generate source-like images. The pseudo-labels output from the middle segmentation network is used to supervise the learning progress of the desired model (bottom segmentation network). In the second stage, the circular learning and pixel-adaptive mask refinement are used to further improve the desired model performance. With this approach, we achieved satisfactory abdominal multi-organ segmentation performance, outperforming the existing state-of-the-art domain adaptation methods. The proposed approach can be easily extended to situations in which target annotation data exist. With only one labeled MR volume, the performance can be leveled with that of supervised learning. Furthermore, the proposed approach is proven to be effective for source-free unsupervised domain adaptation in reverse direction.
Display omitted
•An improved deep neural network for capturing the high-dimensional features of wind-solar energy.•A refined model and a two-stage solution for cascade hydropower stations.•Proof of ...the applicability of only a hydro-wind-solar hybrid system to satisfy power transmission.•Improvement of the quality of generated scenarios helps enhance the hybrid system performance.
The high penetration of variable renewable energy sources (RESs) has greatly increased the difficulty in power system scheduling and operation. To fully utilize the complementary characteristics of various RESs, a stochastic optimization model considering the strong regulation capacity of cascade hydropower stations and the uncertainty of wind and photovoltaic (PV) power is presented. Based on the improved generative adversarial networks, the spatial and temporal correlation characteristics between wind farms and PV plants are accurately captured via measured data. Due to the nonlinear features of the hydroelectric plants, linearization methods are adopted to reformulate the original model into a standard mixed integer linear programming (MILP) formulation. Then, the model is solved with a proposed two-stage approach, in which a heuristic algorithm is used to solve the first-stage unit commitment optimization. The cascade hydraulic connection and time delay of the water flow are established in the second stage to exploit the considerably controllable adjustment capability of hydropower generation. A renewable energy base in southwest China is chosen as a detailed case study. The simulation results reveal the potential of the large-scale application of only a hydro-wind-solar hybrid system to satisfy the power transmission demand with the guidance of the coordinated operation strategy, and the performance of the hybrid system can be further enhanced with high-quality scenarios from the proposed deep neural network.
•We analysed over 320 COVID-19 images and 320 healthy control images.•We proposed an improved CNN to extract individual image-level features.•We proposed to use GCN to extract relation-aware ...representations.•We proposed a DFF technology to combine features from GCN and CNN.•The proposed FCGNet gives better performance than 15 state-of-the-art approaches.
(Aim) COVID-19 is an infectious disease spreading to the world this year. In this study, we plan to develop an artificial intelligence based tool to diagnose on chest CT images.
(Method) On one hand, we extract features from a self-created convolutional neural network (CNN) to learn individual image-level representations. The proposed CNN employed several new techniques such as rank-based average pooling and multiple-way data augmentation. On the other hand, relation-aware representations were learnt from graph convolutional network (GCN). Deep feature fusion (DFF) was developed in this work to fuse individual image-level features and relation-aware features from both GCN and CNN, respectively. The best model was named as FGCNet.
(Results) The experiment first chose the best model from eight proposed network models, and then compared it with 15 state-of-the-art approaches.
(Conclusion) The proposed FGCNet model is effective and gives better performance than all 15 state-of-the-art methods. Thus, our proposed FGCNet model can assist radiologists to rapidly detect COVID-19 from chest CT images.
•We proposed a general ResGNet Framework that is suitable for image classification tasks.•We propose three novel models for COVID-19 detection.•It is the first attempt at applying graph convolutional ...neural network for COVID-19 detection.•Compared to SOTA, our model achieved the best performance in terms of accuracy.
The widely spreading COVID-19 has caused thousands of hundreds of mortalities over the world in the past few months. Early diagnosis of the virus is of great significance for both of infected patients and doctors providing treatments. Chest Computerized tomography (CT) screening is one of the most straightforward techniques to detect pneumonia which was caused by the virus and thus to make the diagnosis. To facilitate the process of diagnosing COVID-19, we therefore developed a graph convolutional neural network ResGNet-C under ResGNet framework to automatically classify lung CT images into normal and confirmed pneumonia caused by COVID-19. In ResGNet-C, two by-products named NNet-C, ResNet101-C that showed high performance on detection of COVID-19 are simultaneously generated as well. Our best model ResGNet-C achieved an averaged accuracy at 0.9662 with an averaged sensitivity at 0.9733 and an averaged specificity at 0.9591 using five cross-validations on the dataset, which is comprised of 296 CT images. To our best knowledge, this is the first attempt at integrating graph knowledge into the COVID-19 classification task. Graphs are constructed according to the Euclidean distance between features extracted by our proposed ResNet101-C and then are encoded with the features to give the prediction results of CT images. Besides the high-performance system, which surpassed all state-of-the-art methods, our proposed graph construction method is simple, transferrable yet quite helpful for improving the performance of classifiers, as can be justified by the experimental results.
Background
Non‐small cell lung cancer (NSCLC) is one of the most common human malignancies and the leading cause of cancer‐related death. Over the past few decades, genomic alterations of cancer ...driver genes have been identified in NSCLC, and molecular testing and targeted therapies have become standard care for lung cancer patients. Here we studied the unique genomic profile of driver genes in Chinese patients with NSCLC by next‐generation sequencing (NGS) assay.
Materials and Methods
A total of 1,200 Chinese patients with NSCLC were enrolled in this study. The median age was 60 years (range: 26–89), and 83% cases were adenocarcinoma. NGS‐based genomic profiling of major lung cancer‐related genes was performed on formalin‐fixed paraffin‐embedded tumor samples and matched blood.
Results
Approximately 73.9% of patients with NSCLC harbored at least one actionable alteration recommended by the National Comprehensive Cancer Network guideline, including epidermal growth factor receptor (EGFR), ALK, ERBB2, MET, BRAF, RET, and ROS1. Twenty‐seven patients (2.2%) harbored inherited germline mutations of cancer susceptibility genes. The frequencies of EGFR genomic alterations (both mutations and amplification) and ALK rearrangement were identified as 50.1% and 7.8% in Chinese NSCLC populations, respectively, and significantly higher than the Western population. Fifty‐six distinct uncommon EGFR mutations other than L858R, exon19del, exon20ins, or T790M were identified in 18.9% of patients with EGFR‐mutant NSCLC. About 7.4% of patients harbored both sensitizing and uncommon mutations, and 11.6% of patients harbored only uncommon EGFR mutations. The uncommon EGFR mutations more frequently combined with the genomic alterations of ALK, CDKN2A, NTRK3, TSC2, and KRAS. In patients <40 years of age, the ALK‐positive percentage was up to 28.2%. Moreover, 3.2% of ALK‐positive patients harbored multi ALK rearrangements, and seven new partner genes were identified.
Conclusion
More unique features of cancer driver genes in Chinese NSCLC were identified by next‐generation sequencing. These findings highlighted that NGS technology is more feasible and necessary than other molecular testing methods, and suggested that the special strategies are needed for drug development and targeted therapy for Chinese patients with NSCLC.
Implications for Practice
Molecular targeted therapy is now the standard first‐line treatment for patients with advanced non‐small cell lung cancer (NSCLC). Samples of 1,200 Chinese patients with NSCLC were analyzed through next‐generation sequencing to characterize the unique feature of uncommon EGFR mutations and ALK fusion. The results showed that 7.4% of EGFR‐mutant patients harbored both sensitizing and uncommon mutations and 11.6% harbored only uncommon mutations. Uncommon EGFR mutations more frequently combined with the genomic alterations of ALK, CDKN2A, NTRK3, TSC2, and KRAS. ALK fusion was more common in younger patients, and the frequency decreased monotonically with age. 3.2% of ALK‐positive patients harbored multi ALK rearrangement, and seven new partner genes were identified.
Lung cancer is the most fatal malignancy in China. This study assessed genomic alterations of driver genes in a cohort of Chinese patients with non‐small cell lung cancer. This article reports the resulting analysis of germline mutations, EGFR variations, and ALK rearrangements, the most common and important driver genes in Chinese NSCLC population.
Till August 17, 2020, COVID-19 has caused 21.59 million confirmed cases in more than 227 countries and territories, and 26 naval ships. Chest CT is an effective way to detect COVID-19. This study ...proposed a novel deep learning model that can diagnose COVID-19 on chest CT more accurately and swiftly. Based on traditional deep convolutional neural network (DCNN) model, we proposed three improvements: (i) We introduced stochastic pooling to replace average pooling and max pooling; (ii) We combined conv layer with batch normalization layer and obtained the conv block (CB); (iii) We combined dropout layer with fully connected layer and obtained the fully connected block (FCB). Our algorithm achieved a sensitivity of 93.28% ± 1.50%, a specificity of 94.00% ± 1.56%, and an accuracy of 93.64% ± 1.42%, in identifying COVID-19 from normal subjects. We proved using stochastic pooling yields better performance than average pooling and max pooling. We compared different structure configurations and proved our 3CB + 2FCB yields the best performance. The proposed model is effective in detecting COVID-19 based on chest CT images.
Objective
To investigate whether machine learning-based analysis of MR radiomics can help improve the performance PI-RADS v2 in clinically relevant prostate cancer (PCa).
Methods
This IRB-approved ...study included 54 patients with PCa undergoing multi-parametric (mp) MRI before prostatectomy. Imaging analysis was performed on 54 tumours, 47 normal peripheral (PZ) and 48 normal transitional (TZ) zone based on histological-radiological correlation. Mp-MRI was scored via PI-RADS, and quantified by measuring radiomic features. Predictive model was developed using a novel support vector machine trained with: (i) radiomics, (ii) PI-RADS scores, (iii) radiomics and PI-RADS scores. Paired comparison was made via ROC analysis.
Results
For PCa versus normal TZ, the model trained with radiomics had a significantly higher area under the ROC curve (Az) (0.955 95% CI 0.923–0.976) than PI-RADS (Az: 0.878 0.834–0.914, p < 0.001). The Az between them was insignificant for PCa versus PZ (0.972 0.945–0.988 vs. 0.940 0.905–0.965, p = 0.097). When radiomics was added, performance of PI-RADS was significantly improved for PCa versus PZ (Az: 0.983 0.960–0.995) and PCa versus TZ (Az: 0.968 0.940–0.985).
Conclusion
Machine learning analysis of MR radiomics can help improve the performance of PI-RADS in clinically relevant PCa.
Key Points
•
Machine
-
based analysis of MR radiomics outperformed in TZ cancer against PI
-
RADS
.
•
Adding MR radiomics significantly improved the performance of PI
-
RADS
.
•
DKI
-
derived Dapp and Kapp were two strong markers for the diagnosis of PCa
.
With the rapid development of 5G/6G and IoV technologies, the multi-temporal variable parametric relationships between devices and unused idle resources at the edge in highly dynamic IoV can affect ...the QoS experience of users. Therefore, we construct a framework for air-ground collaborative offloading and content acquisition model first in edge VANETs environment, where users are able to select compute and cache content resources from different edge nodes for edge services. Secondly, we propose a dynamic orchestration algorithm for knowledge graphs of edge services based on multitemporal variable parametric knowledge relations which reduces the interference of redundant information. Then, a non-convex optimization problem is established by considering the system utility function weighted by multiple performance metrics under environmental constraints and different service policies, and the problem is transformed and analyzed by a theoretical perspective using block coordinate descent and successive convex approximation methods to obtain the theoretical solution. In terms of simulation, we propose a knowledge graph-aware air-ground collaborative offloading and content acquisition SAC algorithm in VANETs (VAKOCS) to obtain the simulation solution of the problem and validate the problem in both directions. The proposed VAKOCS algorithm improves the network reward by 6.5%, 23%, 47% and 72%, respectively. Task offloading delay is reduced by 32%, 63%, 78% and 87%, respectively, and the content rental utility is reduced by 6%, 51% and 72%, respectively, while the error between the simulation solution and the theoretical solution is 0.45.
Compared with traditional remoting image, there is a large amount of spectral information in the hyperspectral image (HSI), which makes HSI better reflect the actual condition of surface features. ...However, due to the limitations of imaging conditions, HSI tends to have a lower spatial resolution. In order to overcome this issue, we propose a spectral-spatial attention-based U-Net named SSAU-Net for HSI and multispectral image (MSI) fusion. The SSAU-Net constructs a spectral-spatial attention module by a coordinate-attention (CA) module and an efficient pyramid split attention (ESPA) module, which can enhance the image's spectral information and spatial information. Meanwhile, the proposed network fully extracts the shallow and deep features of the images, and finally generates high-resolution (HR) hyperspectral images. Compared with state-of-the-art HSI-MSI fusion methods, the experimental results verify that the proposed method has a better subjective and objective fusion effect.