Search

Graph clustering is an essential task in data analysis. Recently, there has been a notable trend in the application of deep learning in graph clustering. However, it is worth noting that existing deep graph clustering methods primarily focus on the topological information of nodes while overlooking the use of attribute information in the learned feature representations. In addition, they do not address the consistency of the space distribution in attribute and structure clustering results. To tackle these critical issues, a novel structured deep graph clustering network with consistency constraint (CC-DGC) is proposed. The network first constructs an autoencoder to learn and transmit hierarchical attribute information to the graph autoencoder (GAE). Subsequently, the GAE integrates the received hierarchical attribute information with extracted topological information to generate enhanced clustering representations. Moreover, this paper designs a consistency constraint module to promote consistency between the autoencoder and the GAE by optimizing the cluster space distributions they produce. Finally, the feature extraction and clustering classification processes are synchronized and optimized in a self-supervised manner within a unified framework. Extensive experiments illustrate that the proposed CC-DGC demonstrates superiority over the state-of-the-art deep graph clustering methods on five benchmark datasets. [ABSTRACT FROM AUTHOR]

Normal mode separation and information extraction are popular topics in marine acoustics. The extraction of modal wavenumbers and modal depth functions using the data of a vertical line array is beneficial to various applications. In this paper, we systematically introduce a normal mode separation method based on the consistency restraint of the horizontal wavenumber, obtain the modal depth functions using the singular value decomposition (SVD) method, determine the vertical distribution of the horizontal wavenumber, and determine the SVD accuracy by the consistency of modal wavenumbers on the vertical direction. Moreover, this method can be used to determine the SVD accuracy without true depth functions. When the degeneracy problem leads to a poor SVD effect, the linear relation between the degenerate and true modes is obtained, and the minimum standard deviation of the vertical distribution of the modal wavenumbers is taken as the loss function, thus optimizing the SVD result and improving the accuracy of modal wavenumber estimation. The proposed method is validated using simulated data. In addition, the modes are successfully extracted from ambient noise.

Currently, most pointcloud classification methods heavily rely on huge numbers of labeled samples. Notably, labeling a large-scale multispectral LiDAR (MS-LiDAR) pointcloud is time-consuming and costly. To address this issue, we propose a feature perturbation weakly supervised network for classifying MS-LiDAR pointclouds using a few labeled samples, termed as FPWS-Net. In the FPWS-Net, we innovatively design a dual semantic inference structure, including a primary semantic inference module and an auxiliary semantic inference module. To provide the network with rich, accurate supervised signals, we embed kernel point convolution (KPConv) into the network for modelling the contextual information of MS-LiDAR pointclouds and propagating the signals between labeled and unlabeled points. Additionally, to constrain feature perturbations resulting from the dual semantic inference structure, fully leverage unlabeled points, and fit the architecture of the FPWS-Net, we combine consistency constraint and mutual pseudo-labeling loss. The proposed FPWS-Net is tested on two datasets, and achieves at least an average F1-score of 83.69 %, an mIoU of 78.81 %, and an OA of 95.97 % using only 0.1 % labeled points. The comparative experimental results demonstrate that the FPWS-Net not only outperforms the state-of-the-art (SOTA) weakly supervised networks, but also achieves the comparable classification performance to the fully supervised methods in the airborne MS-LiDAR pointcloud classification tasks.

Multimodal emotion recognition in conversations（MERC） is a task to predict the emotional category of the discourse in a dialogue based on its textual, audio, and visual modality. Existing studies focus on multimodal feature extraction and fusion of discourse context without fully considering the utilization of emotional features of different speakers. Therefore, this paper proposed a model of multimodal dialogue emotion recognition based on a consistent graph convolutional network. The model first constructed a graph convolutional network of multimodal feature learning and fusion, and obtained the context features of each discourse. On this basis, the average features of the speaker in the complete dialogue as the constraint to make the model learn more reasonable discourse features, so as to improve the performance of predicting emotion class. The paper compared with other baseline models on two benchmark datasets IEMOCAP and MELD. And the results show that the proposed model is superior to the other models. In addition, the paper verifies the consistency constraints and other components of the model through ablation experiments. [ABSTRACT FROM AUTHOR]

Radiation therapy is a fundamental cancer treatment in the clinic. However, to satisfy the clinical requirements, radiologists have to iteratively adjust the radiotherapy plan based on experience, causing it extremely subjective and time-consuming to obtain a clinically acceptable plan. To this end, we introduce a transformer-embedded multi-task dose prediction (TransMTDP) network to automatically predict the dose distribution in radiotherapy. Specifically, to achieve more stable and accurate dose predictions, three highly correlated tasks are included in our TransMTDP network, i.e. a main dose prediction task to provide each pixel with a fine-grained dose value, an auxiliary isodose lines prediction task to produce coarse-grained dose ranges, and an auxiliary gradient prediction task to learn subtle gradient information such as radiation patterns and edges in the dose maps. The three correlated tasks are integrated through a shared encoder, following the multi-task learning strategy. To strengthen the connection of the output layers for different tasks, we further use two additional constraints, i.e. isodose consistency loss and gradient consistency loss, to reinforce the match between the dose distribution features generated by the auxiliary tasks and the main task. Additionally, considering many organs in the human body are symmetrical and the dose maps present abundant global features, we embed the transformer into our framework to capture the long-range dependencies of the dose maps. Evaluated on an in-house rectum cancer dataset and a public head and neck cancer dataset, our method gains superior performance compared with the state-of-the-art ones. Code is available at https://github.com/luuuwen/TransMTDP. [ABSTRACT FROM AUTHOR]

Layer segmentation of Optical Coherence Tomography (OCT) images is an important step in diagnosing retinal diseases. However, the presence of some artifacts and noise in OCT images often leads to unsatisfactory layer segmentation results. Especially when the number of layers to be segmented is particularly large, the boundaries between layers are indistinguishable, which poses a great challenge to automatic and accurate segmentation. To solve these problems, we propose a novel multi-task dual boundary-aware network to improve the retinal layer segmentation performance in OCT images. Specifically, based on the hierarchical relationship between retinal layers, we design a dual boundary representation method to encode the bidirectional boundary information between layers. Then we design a multi-task architecture and a novel consistency loss to utilize the boundary representation to make the segmentation more accurate. For evaluation, we have built a large-scale OCT layer segmentation dataset with 1,200 images. The comprehensive experimental results show that our method achieves superior performance over other state-of-the-art algorithms.

Fine-grained visual classification (FGVC) is much more challenging than traditional classification tasks due to the inherently subtle intra-class object variations. Recent works are mainly part-driven (either explicitly or implicitly), with the assumption that fine-grained information naturally rests within the parts. In this paper, we take a different stance, and show that part operations are not strictly necessary – the key lies with encouraging the network to learn at different granularities and progressively fusing multi-granularity features together. In particular, we propose: (i) a progressive training strategy that effectively fuses features from different granularities, and (ii) a consistent block convolution that encourages the network to learn the category-consistent features at specific granularities. We evaluate on several standard FGVC benchmark datasets, and demonstrate the proposed method consistently outperforms existing alternatives or delivers competitive results. Codes are available at https://github.com/PRIS-CV/PMG-V2. [ABSTRACT FROM AUTHOR]

Despite that deep learning models have achieved remarkable performance in medical image segmentation, their performance is often limited on testing images from new centers with a domain shift. To achieve Domain Generalization (DG) for medical image segmentation, we propose a Domain Composition and Attention-based Network (DCA-Net) combined with structure- and style-based data augmentation that generates unlabeled synthetic images for training. First, DCA-Net represents features in one certain domain by a linear combination of a set of basis representations that are learned by parallel domain preceptors with a divergence constraint. The linear combination is used to calibrate the feature maps of an input image, which enables the model to generalize to unseen domains. Second, considering the number of domains and images for training is limited, we employ generative models to synthesize images with a higher structure diversity, and to leverage the unlabeled synthetic images, we introduce a consistency constraint for their predictions under style augmentation based on frequency amplitude mixture. Additionally, a Test-Time Frequency Augmentation (TTFA) is proposed to neutralize the domain shift from the target to source domains. Experimental results on two multi-domain datasets for fundus structure and nasopharyngeal carcinoma segmentation showed that: (1) our method significantly outperformed several existing DG methods, and (2) the model's generalizability was largely improved by domain composition and attention modules; (3) by leveraging the unlabeled synthetic images and the TTFA, the model could better deal with images from unseen domains. • A Domain Composition and Attention-based Network (DCA-Net) for generalizable medical image segmentation. • Consistency constraint on unlabeled synthesized images is proposed to improve generalizability. • Test-Time Frequency Augmentation (TTFA) reduces the domain gap between source and unseen target domains. • State-of-the-art domain generalization performance was achieved on two multi-domain datasets. [ABSTRACT FROM AUTHOR]

Although novel point cloud semantic segmentation schemes that continuously surpass state-of-the-art results exist, the success of learning an effective model typically relies on the availability of abundant labeled data. However, data annotation is a time-consumng and labor-intensive task, particularly for large-scale airborne laser scanning (ALS) point clouds involving multiple classes in urban areas. Therefore, simultaneously obtaining promising results while significantly reducing labeling is crucial. In this study, we propose a deep-learning-based weakly supervised framework for the semantic segmentation of ALS point clouds. This is to exploit implicit information from unlabeled data subject to incomplete and sparse labels. Entropy regularization is introduced to penalize class overlap in the predictive probability. Additionally, a consistency constraint is designed to improve the robustness of the predictions by minimizing the difference between the current and ensemble predictions. Finally, we propose an online soft pseudo-labeling strategy to create additional supervisory sources in an efficient and nonparametric manner. Extensive experimental analysis using three benchmark datasets demonstrates that our proposed method significantly boosts the classification performance without compromising the computational efficiency, considering the sparse point annotations. It outperforms the current weakly supervised methods and achieves a result comparable to that of full supervision competitors. Considering the ISPRS Vaihingen 3D data, using only 1‰ labels, our method achieved an overall accuracy of 83.0% and an average F1 score of 70.0%. These increased by 6.9% and 12.8%, respectively, compared to the model trained only using sparse label information. [ABSTRACT FROM AUTHOR]

Under differential privacy, to solve the optimal publishing problem with multiple consistency constraints, an approximation method of multiple consistency constraints was proposed by the theoretical analysis of the principle of optimal consistency release.The main idea was to divide the consistency constraint problem into several consistency constraint sub-problems and then achieve the original problem's optimal consistency release by solving each consistency constraint sub-problem repeatedly and independently.The advantage was that after the consistency constraint problem divided, the sub-problems were often easier to solve, or the technology to achieve optimal and consistent release of sub-problems is quite mature.Therefore more complex differential privacy optimal release problem could be solved.After analysis, the approximation method's convergence was fully demonstrated, ensuring that any partition of consistency constrained sub-problems can always achieve the optimal consistency release of the original problem.Furthermore, taking the sales histogram publishing as an example, based on the approximation method of multiple consistency constraints, a parallel algorithm was designed with optimal consistency release under differential privacy.The experimental results show that the algorithm's efficiency is 400 times higher than that of the general solution, and the algorithm can process millions of large-scale data.

Joint registration of point clouds obtained from multiple views is a key step of reconstruction for blades. However, due to the structural and surface characteristics of blades, some views do not meet the overlap constraints of registration, which results in significant initial errors of pose estimation. Thus, we propose a novel approach to recover the accuracy of poses estimation. The proposed method is robust to overlap extent of views through a stochastic framework. The approach formulates a variable-parameters graph optimization problem. Then a simulated annealing algorithm is used to solve the global optimal parameters. The candidate parameters in the simulated annealing process are obtained through the improved unscented Kalman filter, which reduces the initial errors and enhances the information matrices. The acceptance of the candidate parameters is determined by the optimization problem constrained by joint point correspondences and closed-loop consistency. And the parameters that can improve the registration accuracy are selected. We test our algorithm with simulated synthetic data and real data obtained by the robot measurement system. We compare the proposed algorithm with several state-of-the-art algorithms. The experimental results show that in the presence of significant initial errors, our method can estimate the poses more accurately and obtain better blade reconstructions. [ABSTRACT FROM AUTHOR]

Copyright of Journal on Communication / Tongxin Xuebao is the property of Journal on Communications Editorial Office and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Multi-label Chest X-ray (CXR) images often contain rich label relationship information, which is beneficial to improve classification performance. However, because of the intricate relationships among labels, most existing works fail to effectively learn and make full use of the label correlations, resulting in limited classification performance. In this study, we propose a multi-label learning framework that learns and leverages the label correlations to improve multi-label CXR image classification. In this paper, we capture the global label correlations through the self-attention mechanism. Meanwhile, to better utilize label correlations for guiding feature learning, we decompose the image-level features into label-level features. Furthermore, we enhance label-level feature learning in an end-to-end manner by a consistency constraint between global and local label correlations, and a label correlation guided multi-label supervised contrastive loss. To demonstrate the superior performance of our proposed approach, we conduct three times 5-fold cross-validation experiments on the CheXpert dataset. Our approach obtains an average F1 score of 44.6% and an AUC of 76.5%, achieving a 7.7% and 1.3% improvement compared to the state-of-the-art results. More accurate label correlations and full utilization of the learned label correlations help learn more discriminative label-level features. Experimental results demonstrate that our approach achieves exceptionally competitive performance compared to the state-of-the-art algorithms. • An end-to-end multi-label learning framework is proposed for multi-label CXR image classification. • Consistency loss between global and local label correlations help learns more accurate label correlation. • Label correlation guided multi-label supervised contrastive learning help learn more discriminative label-level features. • The proposed model outperforms other multi-label classification methods on the CheXpert dataset. [ABSTRACT FROM AUTHOR]

Copyright of Electric Power Automation Equipment / Dianli Zidonghua Shebei is the property of Electric Power Automation Equipment Press and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

For the low sparse reconstruction efficiency caused by high resolutions and large volumes of UAV (unmanned aerial vehicle) images, this paper proposes an algorithm for decreasing the number of image pairs and improving the efficiency of outlier removal. Firstly, rough POS (positioning and orientation system) is calculated for each image with the use of GNSS/IMU (Global Navigation Satellite System/ inertial measurement unit) data and camera installation angles. Secondly, to reduce image combination complexity, topological connection analysis is used for image pairs selection. Considering high outlier ratios of initial matches, the hierarchical motion consistency constraint (HMCC) is designed to achieve the high efficiency of geometrical verification strategies. The proposed solutions are verified by using four datasets captured with different oblique systems. Results demonstrate that without accuracy sacrifices, the proposed solutions can achieve efficient and reliable reconstruction. [ABSTRACT FROM AUTHOR]

The high performance of state-of-the-art deep learning methods for 3D hand pose estimation heavily depends on a large annotated training set. However, it is difficult and time-consuming to obtain the annotations for 3D hand poses. To leverage unannotated images to reduce the annotation cost, we propose a semi-supervised method based on Multi-Task and Multi-View Consistency (MTMVC) for hand pose estimation. First, we obtain the joints based on heatmap prediction and coordinate regression parallelly and encourage their consistency. Second, we introduce multi-view consistency to encourage the predicted poses to be rotation-invariant. Thirdly, to make the network pay more attention to the hand region, we propose a spatially weighted consistency. Experiments on four public datasets showed that our proposed MTMVC outperformed existing semi-supervised hand pose estimation methods, and by only using half of the annotations, the accuracy of our method was comparable to those of several state-of-the-art fully supervised methods. • A semi-supervised 3D hand pose estimation framework for large-scale hand pose estimation datasets. • A multi-task consistency constraint combining detection-based and regression-based hand pose estimation methods. • A multi-view consistency to encourage the results to be invariant to in-plane transformations. • Our method is comparable to fully supervised methods while reducing the annotation cost by half. [ABSTRACT FROM AUTHOR]

To solve the global optimization problem in heavy machinery, a multidisciplinary design optimization (MDO) computing environment is developed based on high level architecture (HLA). The computing environment encapsulates the analysis optimization tools in relevant disciplinary fields to seek global optimization design based on a single disciplinary optimization. The goal of a minimum discrepancy between the subsystem-level design variables and system-level target values is reached by establishing multidisciplinary consistency constraints. The HLA-compliant optimization model realizes collaboration and interoperation between the disciplines through mapping the coupled state variables to corresponding published or subscribed attributes of an object class according to the rules of HLA. Finally, an example application of designing a hydraulic-powered support is presented to demonstrate the effectiveness of the optimization environment. [ABSTRACT FROM AUTHOR]