Abstract The accurate detection of moving objects is essential in various applications of artificial intelligence, particularly in the field of intelligent surveillance systems. However, the moving cast shadow detection significantly decreases the precision of moving object detection because they share similar motion characteristics. To address the issue, the authors propose an innovative approach to detect moving cast shadows by combining the hybrid feature with a broad learning system (BLS). The approach involves extracting low‐level features from the input and background images based on colour constancy and texture consistency principles that are shown to be highly effective in moving cast shadow detection. The authors then utilise the BLS to create a hybrid feature and BLS uses the extracted low‐level features as input instead of the original data. BLS is an innovative form of deep learning that can map input to feature nodes and further enhance them by enhancement nodes, resulting in more compact features for classification. Finally, the authors develop an efficient and straightforward post‐processing technique to improve the accuracy of moving object detection. To evaluate the effectiveness and generalisation ability, the authors conduct extensive experiments on public ATON‐CVRR and CDnet datasets to verify the superior performance of our method by comparing with representative approaches.

In recent years, device-free passive localization leveraging Wi-Fi channel state information (CSI) has emerged as a prominent technique for indoor positioning, yet the nonlinear interactions and signal superposition among multiple targets, coupled with occlusion and shadowing effects, significantly complicate the localization task, rendering multitarget device-free passive localization a substantial challenge in the field. In this article, we propose a novel device-free passive multitarget indoor localization approach based on multilabel learning (MLL) and unsupervised domain adaptation, denoted as MLDA-MultiLoc. It segments the localization area into multiple training point regions, reformulating the multitarget problem as a multilabel classification task. MLDA-MultiLoc employs a fusion representation model that capitalizes on the spatio-temporal redundancy of CSI amplitude and phase, effectively mapping these features into a unified representation domain. This model is optimized to enhance the discriminative power of the fusion fingerprint (HDFF) by maximizing spatial metrics. Acknowledging the nonlinear influence of multiple targets on CSI, MLDA-MultiLoc incorporates a fusion generation network to synthesize multitarget fingerprints from multiple single-target fingerprints, creating virtual samples for multitarget scenarios. This process facilitates the training of a deep learning-based multilabel classifier, leveraging MLL for robust parameter optimization. Furthermore, MLDA-MultiLoc introduces an unsupervised domain adaptation technique that utilizes a meta-learning dual-stream structure. This method effectively bridges the gap between virtual and real fingerprint samples, ensuring accurate multitarget localization in complex, dynamic indoor settings. Extensive experiments have confirmed the superiority of MLDA-MultiLoc over existing state-of-the-art systems, showcasing its effectiveness in real-world indoor environments.

The Segment Anything Model (SAM) excels in general segmentation but encounters difficulties in medical imaging due to few-shot learning challenges, particularly with extremely limited annotated data. Existing approaches often suffer from insufficient feature extraction and inadequate loss function balancing, resulting in decreased accuracy and poor generalization. To address these issues, we propose BiASAM, which uniquely incorporates two bidirectional attention mechanisms into SAM for medical image segmentation. Firstly, BiASAM integrates a spatial-frequency attention module to improve feature extraction, enhancing the model's ability to capture both fine and coarse details. Secondly, we employ an attention-based gradient update mechanism that dynamically adjusts loss weights, boosting the model's learning efficiency and adaptability in data-scarce scenarios. Additionally, BiASAM utilizes the point and box fusion prompt to enhance segmentation precision at both global and local levels. Experiments across various medical datasets show BiASAM achieves performance comparable to fully supervised methods with just two labeled samples.

Multiview clustering (MVC) with contrastive learning (CL) has attracted considerable interest. Nevertheless, current methods have specific drawbacks since the coherence between views in them is limited either at the feature representation level or the cluster representation level. Besides, certain methods demonstrate subpar performance and limited robustness when handling noisy data. This article introduces an efficient multilevel fusion CL framework for MVC called EMLFCL. The EMLFCL model seamlessly incorporates a shared multi-layer perceptron (MLP) network (MNet) and a fusion network (FNet) to capture and merge common representation information, which effectively eliminates the impact of view-specific private information during the clustering process. Specifically, we establish an efficient multilevel CL strategy at both the feature representation level and the clustering representation level. Rather than rely on pairwise comparisons between views, our proposed CL strategy makes comparisons between different views and the anchor view. Since the anchor view contains abundant shared information, this strategy effectively mitigates the influence of view-specific and noisy view information on model performance. The proposed method outperforms numerous advanced approaches, as evidenced by extensive experiments conducted on eleven challenging multiview datasets. Particularly, it achieves 66.4%, 74.7%, 82.3%, and 86.4% clustering accuracies on the four Caltech datasets with different views, respectively.

Deep learning exhibits powerful capability in image inpainting task, particularly in generating pixel-level details closely with the human visual perception. However, the complex background or larger missing regions make it still encounters the artifacts. Many researchers have investigated that prior information is crucial for guiding the image inpainting. In this article, we introduce the dual-attention mechanism, including lightweight spatial attention and linearized attention, to construct an end-to-end texture and structure-guided image inpainting method. In the first stage, we build the detail inpainting network with the lightweight spatial attention. In this model, the extracted texture and structural features are fused with multi-layers and then the fused detail image is considered as the prior to guide the detail repair of corrupted images. In the second stage, we construct the content completing network by the repaired detail and the linearized Transformer module. This module not only overcomes the limitation of the receptive field size of convolutional kernels that can improve the long-range modeling of features but also can significantly reduce the computational complexity of the original Transformer. To demonstrate the superior effectiveness of the proposed method, we perform extensive experiments with advanced models on three datasets: CelebA-HQ, Places2, and Paris Street Views. Comparative results manifest that our method achieves excellent image inpainting results that are conform to the human visual system. The code is available at https://github.com/QinLab-WFU/TSGDAM

On event-based social platforms such as Meetup and Douban, online groups serve as more than virtual communities for users to share experiences, they also provide an essential pathway for users to discover and participate in offline events. As the number of groups grows, it imposes the need of the study of online group recommendation. Despite there being many existing approaches to solve this problem, they all ignore the phenomenon that the groups that users participate in often contain a number of similar users. This phenomenon implies that similar users play a crucial role in identifying the groups that users are likely to join. In order to exploit similar users to improve the recommendation performance, we propose an effective multi-scale contrastive learning system for online Group Recommendation services, which is with a two-Tower model in event-based social networks (Tower4GR). Specifically, we first adopt the two-tower model to capture the interactive signals within the sequences and groups. We then incorporate the features of similar users into the sequence encoder, and aggregate the relevant users’ features into the group encoder, through which the preferred groups of similar users are more likely to be discovered by the target user. Finally, we propose an effective multi-scale contrastive learning framework for the two-tower architecture. It derives self-supervision signals from both same-scale data and cross-scale data, thereby extracting more meaningful data patterns. Moreover, the framework strengthens the cooperative associations between two towers. Extensive experiments on three real-world datasets from Meetup demonstrate the superiority of our proposed model over existing state-of-the-art models.

Source-Free Domain Adaptation (SFDA) aims to adapt a pre-trained model to unlabeled target domain data without access to source domain data, presenting a significant challenge for medical image segmentation. Most current approaches address this challenge through self-training, employing manually augmented target domain images and pseudo-labels to enforce consistency regularization. However, these approaches still encounter two primary issues. Firstly, manually augmented consistency self-training results in performance degradation due to the semantic mismatch between the target domain images and noisy pseudo-labels. Secondly, they fail to fully exploit the informative content present in the target domain, exhibiting inadequate adaptability, particularly in significant domain gaps. To address these, we introduce the Diffusion-driven Dual-flow SFDA (D2SFDA), the pioneering framework to integrate a diffusion model into SFDA for medical image segmentation. Our D2SFDA framework comprises two novel components: the Diffusion Perturbation Flow (DPF) and the Twin-Knowledge Investigation Flow (TKIF). DPF utilizes pseudo-labels to generate diverse and semantically consistent diffusion views, providing more realistic supervision, potentially enhancing model stability. Surprisingly, DPF using only diffusion images outperforms self-training using real images, as evidenced by the superior average Dice score on the BASE1 target domain of the RIGA+ dataset (90.31% vs. 85.64%). Additionally, TKIF rigorously analyzes the target domain with dual-focus consistency regularization on domain-invariant and target domain-specific knowledge, effectively reducing domain gaps, resulting in an improvement from 90.31% to 91.79%. Extensive experiments on two cross-domain datasets confirm that our D2SFDA surpasses state-of-the-art SFDA approaches in effectively addressing domain shift issues. The code is available at https://github.com/M4cheal/D2SFDA.

In recent years, the rise of location-based service applications such as cashier-less shopping, mobile advertisement targeting, and geo-based augmented reality (AR) has been remarkable. These applications offer convenient and interactive experiences by utilizing indoor localization technology. One popular research area in indoor localization is passive fingerprinting localization based on Channel State Information (CSI), which uses general-purpose Wi-Fi platforms and “unconscious cooperative sensing” to achieve device-free localization. However, existing studies face challenges related to inadequate fingerprint richness, limited distinguishability, and inconsistent fingerprint features in real-world dynamic environments. To address these challenges, we prpose MFFLoc in this paper. MFFLoc extracts and processes amplitude and phase information from CSI in a 2D manner. It then fuses the amplitude and phase information using multimodal fusion representation, resulting in rich and distinguishable fused fingerprint features. This approach allows MFFLoc to achieve satisfactory accuracy with just one communication link, reducing deployment costs. To overcome the issue of inconsistent fingerprint features in dynamic environments, MFFLoc proposes an unsupervised domain adaptation method. It employs a dual-flow structure, with one flow operating in the source domain and the other in the target domain. The adaptation layer, with correlated weights, remains unshared between the two flows. Meta-learning is also used to automatically determine the most suitable adaptation layer. Through extensive 6-day experiments conducted in a dynamic indoor environment, MFFLoc showcases superior performance compared to state-of-the-art systems. It demonstrates higher localization accuracy and robustness, making it a promising solution for indoor localization applications.

Segmentation of the Optic Disc (OD) and Optic Cup (OC) is crucial for the early detection and treatment of glaucoma. Despite the strides made in deep neural networks, incorporating trained segmentation models for clinical application remains challenging due to domain shifts arising from disparities in fundus images across different healthcare institutions. To tackle this challenge, this study introduces an innovative unsupervised domain adaptation technique called Multi-scale Adaptive Adversarial Learning (MAAL), which consists of three key components. The Multi-scale Wasserstein Patch Discriminator (MWPD) module is designed to extract domain-specific features at multiple scales, enhancing domain classification performance and offering valuable guidance for the segmentation network. To further enhance model generalizability and explore domain-invariant features, we introduce the Adaptive Weighted Domain Constraint (AWDC) module. During training, this module dynamically assigns varying weights to different scales, allowing the model to adaptively focus on informative features. Furthermore, the Pixel-level Feature Enhancement (PFE) module enhances low-level features extracted at shallow network layers by incorporating refined high-level features. This integration ensures the preservation of domain-invariant information, effectively addressing domain variation and mitigating the loss of global features. Two publicly accessible fundus image databases are employed to demonstrate the effectiveness of our MAAL method in mitigating model degradation and improving segmentation performance. The achieved results outperform current state-of-the-art (SOTA) methods in both OD and OC segmentation.

In recent years, indoor localization has attracted a lot of interest and has become one of the key topics of Internet of Things (IoT) research, presenting a wide range of application scenarios. With the advantages of ubiquitous universal Wi-Fi platforms and the “unconscious collaborative sensing” in the monitored target, Channel State Information (CSI)-based device-free passive indoor fingerprinting localization has become a popular research topic. However, most existing studies have encountered the difficult issues of high deployment labor costs and degradation of localization accuracy due to fingerprint variations in real-world dynamic environments. In this paper, we propose BSWCLoc, a device-free passive fingerprint localization scheme based on the beyond-sharing-weights approach. BSWCLoc uses the calibrated CSI phases, which are more sensitive to the target location, as localization features and performs feature processing from a two-dimensional perspective to ultimately obtain rich fingerprint information. This allows BSWLoc to achieve satisfactory accuracy with only one communication link, significantly reducing deployment consumption. In addition, a beyond-sharing-weights (BSW) method for domain adaptation is developed in BSWCLoc to address the problem of changing CSI in dynamic environments, which results in reduced localization performance. The BSW method proposes a dual-flow structure, where one flow runs in the source domain and the other in the target domain, with correlated but not shared weights in the adaptation layer. BSWCLoc greatly exceeds the state-of-the-art in terms of positioning accuracy and robustness, according to an extensive study in the dynamic indoor environment over 6 days.

In recent years, applying smart contract to Blockchain-based Service Networks (BSNs) has been considered as one of the most promising solution to boost the integration and adoption of Blockchain in big businesses. However, smart contract are especially vulnerable to attack due to poor coding. Although many existing vulnerability detection tools are restricted by rigorous rules that are defined by the experts in advance, these tools are observed to have a high false positive rate in practice. Thus we propose a vulnerability detection framework for smart contract based on the attention mechanism and data flow. The code of smart contract is transformed to a data flow according to the abstract syntax tree that is built from the code. The data flow we built with smart contract code could represent the relationships of code semantic logic. Source code, data flow, and the tags of smart contract code are used as datasets to mask processing. Then, we construct a bidirectional multi-layer transformer architecture based on the attention mechanism to train our dataset. After training, we can get the label of whether there is a vulnerability in the final smart contract. Finally, the model we proposed reaches state-of-the-art results in the practical experiments of smart contract vulnerability detection with 92.54%, 81.79%, and 86.84% in the results Accuracy, Recall, and F1score, respectively.

Recently, end-to-end deep neural networks-based glaucoma diagnosis approaches have been gaining much attention. However, the feature extractor and classier in these approaches are trained together, which is known as coadaptation. Therefore, the feature distribution in them should adapt to particular decision boundaries. To learn generic data representations and improve the generalization ability of the model, this paper designs a three-stage framework for glaucoma diagnosis. In the first stage, preprocessing is utilized to extract the Region of Interesting around the Optic Disc to reduce the computational cost and nonobjective interference. In the second stage, Deep Sparse Autoencoder is designed to learn hybrid features between the deep features and the original features, which could improve the effectiveness of final high-level feature expression. Meanwhile, L1 regularization is introduced and applied on the hybrid features to obtain deep features with high complementarity under small sample problem. In the third stage, the obtained generic feature representations are fed into different classifiers, in which Support Vector Machine classifier achieves the best diagnosis performance. The proposed approach is evaluated on two publicly available databases. Extensive experimental results indicate that our approach outperforms the state-of-the-art approaches with the accuracy of 96.00%, 97.00% and Area Under Curve of 96.94%, 98.28% for REFUGE and Drishti-GS1 databases, respectively.

As one of the most effective feature learning methods, Nonnegative Matrix Factorization (NMF) has been widely used in many scientific fields, such as computer vision, data mining, and bioinformatics. However, NMF is an unsupervised method that cannot fully utilize the label information of data. Thus, its performance is limited in some recognition and classification problems. To remedy this shortcoming, this paper proposes a Semisupervised Discriminative NMF (SDNMF) method. First, we design a Soft-Labeled NMF (SLNMF) model by introducing a soft-label matrix-based regression term into the original NMF, so that the relationship between the soft-label matrix and low-dimensional features can be constructed to improve the discriminative ability of low-dimensional features. Second, to effectively estimate the soft-label matrix, a Label Propagation (LP) model is adopted to fully explore the spatial distribution relationship between the labeled and unlabeled samples. Third, an Adaptive Graph Learning (AGL) model is proposed to exploit the geometric relationship of samples well, which could enhance the performance of LP. Finally, the above three models (i.e., SLNMF, LP, and AGL) are integrated into a unified framework for effective feature learning, which can not only effectively explore the structural relationship matrix between data, but also predict the labels for unknown samples. Moreover, an iterative optimization algorithm is presented to solve our objective function. The convergence and computational complexity analysis of the proposed SDNMF method are also provided. Extensive experiments are conducted on several standard data sets. Compared with related methods, the experimental results verify that the proposed SDNMF method achieves better performance.

Glaucoma is an eye disease that leads to irreversible vision loss. Accurate Optic Disc (OD) and Optic Cup (OC) segmentation can effectively facilitate ophthalmologist in glaucoma diagnosis. Recently, a series of deep learning approaches attain promising performance in OD and OC segmentation but still face the challenge to precisely segment OC boundary with enhanced computational efficiency. To address this issue, we propose a novel network named Robust Multiscale Feature Extraction with Depthwise Separable Convolution (RMSDSC-Net), which can better solve the challenging tradeoff between segmentation performance and network cost. The proposed RMSDSC-Net is mainly composed of Multiscale Input (MSI), Depthwise Separable Convolution Unit (DSCU), Dilated Convolution Block (DCB), and External Residual Connection (ERC). First, the introduction of MSI can reduce the information loss due to the pooling layers used in the network for capturing rich feature representations. Next, to enhance segmentation performance and computational efficiency, this paper designs DSCU and DCB modules to avoid spatial information loss from minor details of the image and preserve more high-level semantic features. Finally, this paper develops ERC established between the encoding layers and decoding layers to minimize the feature degradation problem. Hence, a high segmentation performance can be achieved using a shallow network. To evaluate the performance of the proposed network, extensive experiments have been enforced on two publicly available databases, DRISHTI-GS and REFUGE. Our approach outperforms the state-of-the-art approaches with the Dice Coefficient of (0.978, 0.919) and (0.965, 0.910) for OD and OC segmentation on DRISHTI-GS and REFUGE databases, respectively. As a result, the proposed approach has a strong potential in analyzing fundus images for glaucoma diagnosis.

Retinal vessel segmentation is a vital step for the diagnosis of many early eye-related diseases. In this work, we propose a new deep learning model, namely Channel Attention Residual U-Net (CAR-UNet), to accurately segment retinal vascular and non-vascular pixels. In this model, we introduced a novel Modified Efficient Channel Attention (MECA) to enhance the discriminative ability of the network by considering the interdependence between feature maps. On the one hand, we apply MECA to the "skip connections" in the traditional U-shaped networks, instead of simply copying the feature maps of the contracting path to the corresponding expansive path. On the other hand, we propose a Channel Attention Double Residual Block (CADRB), which integrates MECA into a residual structure as a core structure to construct the proposed CAR-UNet. The results show that our proposed CAR-UNet has reached the state-of-the-art performance on three publicly available retinal vessel datasets: DRIVE, CHASE DB1 and STARE.

Feature representation learning is a key issue in artificial intelligence research. Multiview multimedia data can provide rich information, which makes feature representation become one of the current research hotspots in data analysis. Recently, a large number of multiview data feature representation methods have been proposed, among which matrix factorization shows the excellent performance. Therefore, we propose an adaptive‐weighted multiview deep basis matrix factorization (AMDBMF) method that integrates matrix factorization, deep learning, and view fusion together. Specifically, we first perform deep basis matrix factorization on data of each view. Then, all views are integrated to complete the procedure of multiview feature learning. Finally, we propose an adaptive weighting strategy to fuse the low‐dimensional features of each view so that a unified feature representation can be obtained for multiview multimedia data. We also design an iterative update algorithm to optimize the objective function and justify the convergence of the optimization algorithm through numerical experiments. We conducted clustering experiments on five multiview multimedia datasets and compare the proposed method with several excellent current methods. The experimental results demonstrate that the clustering performance of the proposed method is better than those of the other comparison methods.

With the rapid growth of 6G communication and smart sensor technology, the Internet of Things (IoT) has attracted much attention now. In the 6G‐based IoT applications on the multiprocessor platform, the partitioned scheduling has been widely applied. However, these partitioned scheduling approaches could cause system resource waste and uneven workload among processors. In this paper, a smart semipartitioned scheduling strategy (SSPS) was proposed for mixed‐criticality systems (MCS) in 6G‐based edge computing. Besides tasks’ acceptance rate and weighted schedulability, QoS is considered in SSPS to improve the service quality of the system. The SSPS allocates tasks into each processor, and some tasks can migrate to other processors as soon as possible. By comparing with the several existing algorithms, the experimental results show that the SSPS achieves the best in the schedulability and QoS of the system.

The automatic generation of chest X-ray report is a hot research topic at present. Considering the lack of research on Chinese report generation, we propose a method suitable for lung description in Chinese reports-a model that combines semantic labels and hierarchical LSTM. The model analyzes the anomaly report, extracts high-frequency keywords as semantic labels, and adds the abnormal binary classification module in the encoder to correct the results of the semantic labels for the templated characteristics of the Chinese report. In the design of the decoder, to address the problem of lack of correlation between semantic Labels, a two-layer LSTM model that fuses semantic tags and image features is proposed. The comparison with the baseline experiment shows that the proposed model can effectively improve the quality of report generation.

Retinal vessel segmentation plays an imaportant role in the field of retinal image analysis because changes in retinal vascular structure can aid in the diagnosis of diseases such as hypertension and diabetes. In recent research, numerous successful segmentation methods for fundus images have been proposed. But for other retinal imaging modalities, more research is needed to explore vascular extraction. In this work, we propose an efficient method to segment blood vessels in Scanning Laser Ophthalmoscopy (SLO) retinal images. Inspired by U-Net, "feature map reuse" and residual learning, we propose a deep dense residual network structure called DRNet. In DRNet, feature maps of previous blocks are adaptively aggregated into subsequent layers as input, which not only facilitates spatial reconstruction, but also learns more efficiently due to more stable gradients. Furthermore, we introduce DropBlock to alleviate the overfitting problem of the network. We train and test this model on the recent SLO public dataset. The results show that our method achieves the state-of-the-art performance even without data augmentation.

The precise segmentation of retinal blood vessels is of great significance for early diagnosis of eye-related diseases such as diabetes and hypertension. In this work, we propose a lightweight network named Spatial Attention U-Net (SA-UNet) that does not require thousands of annotated training samples and can be utilized in a data augmentation manner to use the available annotated samples more efficiently. SA-UNet introduces a spatial attention module which infers the attention map along the spatial dimension, and multiplies the attention map by the input feature map for adaptive feature refinement. In addition, the proposed network employs structured dropout convolutional blocks instead of the original convolutional blocks of U-Net to prevent the network from overfitting. We evaluate SA-UNet based on two benchmark retinal datasets: the Vascular Extraction (DRIVE) dataset and the Child Heart and Health Study (CHASE_DB1) dataset. The results show that the proposed SA-UNet achieves state-of-the-art performance on both datasets. The implementation and the trained networks are available on Github1.

Non-negative matrix factorization (NMF) has recently attracted much attention due to its good interpretation in perception science and widely applications in various fields. In this paper, a novel graph regularized NMF algorithm called NMF with locality constrained adaptive graph (NMF-LCAG) is proposed. Compared with other NMF based algorithms, the proposed NMF-LCAG algorithm has the following advantages: 1) Unlike the traditional NMF method which neglects the geometric information of original data, the proposed algorithm introduces a locality constrained graph to discover the latent manifold structure of the data and 2) Different from most graph regularized NMF algorithms in which the graphs are predefined and kept unchanged during the NMF procedure, two locality constraint terms are employed in our NMF-LCAG to adaptively optimize the graph. Thus, the weight matrix of graph and low dimensional features of data can be simultaneously learned by our algorithm, which makes NMF-LCAG more flexible than other approaches. Moreover, an iterative updating strategy is developed to optimize the objective function of our algorithm and the convergence analysis is also given. Extensive experiments are conducted on four face image databases and three UCI datasets to demonstrate the effectiveness of the proposed NMF-LCAG algorithm. Compared with some other related algorithms, the proposed NMF-LCAG algorithm can achieve at least 1% ~ 3% accuracy improvement in most cases.

At present, artificial visual diagnosis of fundus diseases has low manual reading efficiency and strong subjectivity, which easily causes false and missed detections. Automatic segmentation of retinal blood vessels in fundus images is very effective for early diagnosis of diseases such as the hypertension and diabetes. In this paper, we utilize the U-shaped structure to exploit the local features of the retinal vessels and perform retinal vessel segmentation in an end-to-end manner. Inspired by the recently DropBlock, we propose a new method called Structured Dropout U-Net (SD-Unet), which abandons the traditional dropout for convolutional layers, and applies the structured dropout to regularize U-Net. Compared to the state-of-the-art methods, we demonstrate the superior performance of the proposed approach.

With the rapid growing advancement of animation technologies, 3D animated meshes are becoming one of the major data in the industry such as virtual reality. However, treating the animated mesh data efficiently remains a challenging task due to its large scale and limited feature descriptors. In this paper, we present an evolutionary signature for animated meshes based on tempo-spatial segmentation. In specific, we first conduct temporal segmentation to a given animated meshes with sub-motions, then apply spatial segmentation within each temporal segment, and intersect spatial segmentation result for over segmentation. Thirdly, we represent the segmentation results into graphs. Finally, we devise an edge evolution matrix based on the dynamic behaviour of each edge for the evolutionary signature of the input animated mesh. Our experimental results on similarity measurement by using the proposed signature reflect the effectiveness of our method.

Summary Multipath TCP (MPTCP) is a promising transport protocol that allows a multihomed device to simultaneously use multiple network interfaces to send application data over multiple paths. However, although applying MPTCP to data delivery introduces many and attractive benefits, the MPTCP is vulnerable to network attacks. When a path within the MPTCP connection suffers from some types of attacks (eg, a denial‐of‐service attack) and becomes underperforming, it will undoubtedly cause transmission interruption in the stable paths and thus degrade the application‐level performance. Unfortunately, the MPTCP path management mechanism is very simple and cannot timely prevent the usage of underperforming paths in multipath transmission. In this paper, we introduce a new “potentially underperforming” (PU) concept to MPTCP and propose a novel PU‐aware path usage management mechanism ((PU)2M2) for MPTCP aiming to (1) detect and declare an underperforming path and prevent the usage of underperforming paths in multipath transmission, (2) provide a finite‐state‐machine model to change per‐path's state accordingly and effectively manage multiple paths for data transmission, and (3) alleviate the packet reordering problem and make MPTCP avoid throughput performance degradation during network underperforming. We demonstrate the benefits of applying (PU)2M2 to MPTCP.

One major concern when applying Multipath TCP (MPTCP) to the real-time multimedia applications is related to MPTCP's fully-reliable and fully-ordered service nature, which will inevitably degrade users' Quality of Experience (QoE) for multimedia streaming services in a heterogeneous wireless network environment because asymmetric wireless links are commonly with different transmission characteristics and sensitive to variations. In this paper, we first discuss the design considerations of partially reliable-MPTCP associated with the real-time constraint of multimedia streaming. Then we propose a context-aware QoE-oriented MPTCP Partial Reliability extension (PR-MPTCP+) for providing partially reliable multimedia streaming service to an upper layer protocol. Finally, we evaluate the proposed PR-MPTCP+solution using a wide range of multimedia quality metrics.