Federated recommendation systems (FRSs) have recently gained widespread attention due to their ability to train collaborative recommendation models without exchanging raw user data. However, existing FRSs face a severe challenge of data sparsity, which manifests at both the user and item levels. First, user data sparsity: some users may only have a small number of interactions with items, struggling to adequately train the personalized user embedding locally. Second, item data sparsity: some items may only receive a small number of user ratings, causing the global model to lack knowledge about them. Considering these, we propose the Knowledge Enhanced Federated Recommendation System named as KE-FedRS, of which the core idea is to enhance the knowledge of users with few interactions and items with few ratings at both the local and global levels. Specifically, at the local level, we introduce an auxiliary user embedding and average and aggregate this auxiliary embedding across similar users, thereby enriching the knowledge of the local user embedding. At the global level, we propose a hybrid client selection strategy based on item embedding discrepancies, prioritizing clients that exhibit greater divergence in item embeddings from others, thus enhancing the knowledge of items with fewer interactions in the global model. We conduct comprehensive experiments on four real-world datasets, and the results show that the proposed method consistently outperforms baseline approaches in terms of HR@10 and NDCG@10.

The rapid development of the Internet of Things (IoT) and Cyber-Physical Systems (CPS) has led to the rise of Mobile Edge Computing (MEC), enabling low-latency task offloading in dynamic environments. However, existing offloading strategies struggle to generalize across diverse and evolving network topologies, often requiring retraining or fine-tuning when deployed in new scenarios. To address these challenges, we propose AtALT, a transferable task offloading framework that achieves zero-shot transferability. The acronym AtALT is derived from the key components of our method:Attention-AuxiliaryLearning-Transferable task offloading. AtALT integrates an attention-based encoder and an auxiliary learning module. The attention-based encoder dynamically computes topology-agnostic compatibility scores, allowing for flexible task offloading decisions across different network configurations. The auxiliary learning module predicts node states, regularizing the policy learning process and enhancing generalization. Experimental results demonstrate that AtALT outperforms existing methods in transferability and efficiency, making it suitable for deployment in previously unseen environments without the need for further training.

The field of recommender systems aims to predict users’ latent interests by analyzing their preferences and behaviors. However, privacy concerns about user data collection lead to challenges such as incomplete initial information and data sparsity. Federated learning has emerged to address these privacy issues in recommender systems. However, federated recommender systems face heterogeneity among edge devices regarding data features and sample sizes. Moreover, differences in computational and storage capabilities introduce communication overhead and processing delays during parameter aggregation at the third-party server. This article introduces a framework named FedRL , a reinforcement learning federated recommender system for efficient communication using Reinforcement Selector and Hypernet Generator, to address the proposed issues. The Reinforcement Selector dynamically selects participating edge devices and helps to maximize their use of local data resources. Meanwhile, Hypernet Generator optimizes communication bandwidth consumption during the federated learning parameter transmission, enabling rapid deployment and updates of new model architectures or hyperparameters. Furthermore, the framework incorporates item attributes as content embeddings in edge devices’ recommender models, enriching them with global information. Real-world dataset experiments demonstrate that the proposed solution balances recommender quality and communication efficiency. The code for this work is publicly available on GitHub: https://github.com/diyicheng/FedRL .

Accurate drug-target affinity (DTA) prediction holds significant potential in the field of artificial intelligence (AI)-based drug discovery. However, existing methods primarily operate at a single scale, specifically at the macro (residue) scale for target proteins and the micro (atom) scale for drugs, which limits their ability to provide information at micro (atom) scale for targets and macro (functional group, FG) scale for drugs. This limitation hinders a comprehensive understanding of the binding patterns and properties of drug-target pairs. In this paper, we propose a progressive Macro-to-Micro 3D Modeling Network (M²N) that enables macro (residue/FG) to micro (atom) scale unified modeling, termed cross-scale, to predict DTA. Specifically, M²N operates drugs by learning their chemical properties and structural characteristics from a 3D FG graph to a 3D atom graph. Correspondingly, M²N encodes proteins from a 3D residue graph to a 3D atom graph to exploit their sequence, evolutionary, and geometric representations. Such cross-scale 3D modeling scheme allows for coarse-to-fine embedding optimization, followed by an adaptive fusion module to dynamically integrate the refined features by end-to-end learning. Extensive experiments on two datasets indicate that M²N not only outperforms state-of-the-art methods under various conditions, but also provides a new paradigm for target and drug unified modeling.

In the context of rapidly evolving electronic technology, numerous consumer electronic products have begun to employ recommendation systems to enhance user experience. Traditional recommendation systems utilize deep learning to predict user ratings for items; however, this approach requires users to share their data, leading to potential distrust in recommendations. The integration of federated learning into recommendation systems can achieve trustworthy recommendations, but current federated recommendation models necessitate multiple instances of user-item interaction data to learn global parameters. Therefore, this paper introduces Trustworthy Recommendation for Consumer Electronics Using Hypernetworks (TRCE) to ensure trustworthy recommendations for consumer electronics while also catering to users’ personalized needs. Initially, hypernetworks are used to rapidly initialize the recommendation model on the client side, with user preferences embedded as inputs to the hypernetwork to obtain personalized preferences; subsequently, within the client-side recommendation model, Item attribute content embeddings function as global information to offer more contextual facts; finally, attention residual blocks are employed to learn the significance of different item attributes. Experiments demonstrate that this method exhibits commendable recommendation performance on the Movielens1M, Hetrec-movielens, and Douban datasets compared to other models, with improvements in MAE, RMSE, and Accuracy of approximately 4.31%, 4.01%, and 3.70%, respectively.

Personalized sequential recommender has become a key task in the consumer electronics domain. Existing methods for personalized sequential recommenders primarily focus on modeling user behavior and have achieved satisfactory recommender results. However, the inherent quadratic computational complexity in most existing methods often causes typically results in inefficiencies, impeding real-time suggestions. Additionally, these methods cannot be fine-tuned to the personalized needs of users across different scenarios. To address these challenges, we propose the Fine-Grained Global Modeling Learning for Personalized Federated Sequential Recommender (FedSR). Specifically, we design the Associative Mamba Block to model user profiles from a global perspective and enhance prediction efficiency. Furthermore, we introduce the Variable Response Mechanism to fine-tune parameters according to the personalized needs of users. Additionally, we design the Dynamic Magnitude Loss to retain more local personalized information during training. Extensive experiments on three real-world datasets confirms that the proposed FedSR surpasses current approaches in terms of both performance and efficiency, up to 9.48% improvement.

As communication networks and smart gadgets evolve, researchers are becoming increasingly interested in recommender systems. Accurate click-through rate (CTR) prediction improves the performance of recommender systems. However, most current CTR prediction methods have problems in obtaining multi-level feature representations from user input, resulting in biased prediction outputs. Furthermore, CTR prediction models are frequently large-scale deep models, which limits their operational efficiency. To overcome these difficulties, this work introduces the Global Perception Federated Recommender System for Click-Through Rate Prediction (GPFed). The Global Perception Module, in particular, emphasizes the value of various field embeddings from a global viewpoint, focusing on the most salient intra-class features to improve multi-level feature representations in user data. Second, the Compact Tuning Module uses inner products to reduce model size and compression layers to minimize model parameters, resulting in increased operating efficiency. Furthermore, Device-Level Privacy Protection protects device privacy throughout the federated learning process. Experiments on three public datasets reveal that GPFed performs better and more efficiently. Compared to the best baseline models, GPFed improves performance by 10.85%, 3.72%, and 4.74% on the Criteo, Avazu, and MovieLens datasets, respectively.

While metabolic imaging can facilitate early diagnosis by revealing physiological changes of lesions, it is limited by high cost, high radiation risk, and potential renal impairment. Thus, developing an effective approach for Anatomical-to-Metabolic Image Synthesis (A2MIS) is highly required. However, existing methods are heavily hindered by the gap between distinct domains, and fail to provide a confidence score for the synthesized images, severely restricting their clinical applications. Here, we propose a novel Residual-guided Latent Brownian-bridge Co-Diffusion (RLBCD) model for A2MIS. Specifically, RLBCD starts with a co-diffusion process that leverages a residual diffusion branch to capture inter-domain differences, which are injected into an enhanced diffusion branch to maximally reconstruct modality-specific details. Furthermore, to explore desired residual guidance, we investigate the encoder and decoder features in diffusion models, and accordingly design a Hybrid-Granularity Fusion to integrate consistent semantics and complementary information for fine-grained reconstruction. Additionally, a latent consistency score is developed to enhance the restoration of modality-specific information, which also serves as an indicator of the inherent confidence of the synthesized images. Extensive experiments conducted on five public and in-house datasets demonstrate that RLBCD not only outperforms state-of-the-art methods for A2MIS, but also is valuable for downstream clinic applications.

Cross-domain recommendation seeks to utilize insights from several domains to mitigate prevalent challenges in conventional recommendation systems, including data sparsity and cold-start concerns. Previous research has primarily focused on recommendations within the same domain or across different domains, but real-world scenarios often require addressing both simultaneously to meet practical needs. Moreover, most current research rarely addresses data security concerns in recommendation systems, complicating the achievement of a balance between optimal suggestion efficacy and privacy safeguarding. In this paper, we propose a novel cross-domain recommendation framework named PrivRec, which not only addresses both single-domain and multi-domain recommendation scenarios but also emphasizes data security. First, we employ an anomaly detection module to detect and remove abnormal data from the interaction data, thereby enhancing the reliability of the data. We implement a differential privacy mechanism to strengthen the security and confidentiality of user data. In addition, we utilize an innovative contrastive loss function to improve the learning of user preferences across several domains. Furthermore, we introduced a masking mechanism to replace the traditional interaction mask, generating negative samples by randomly selecting items with which users have not interacted, thereby enriching the training dataset. Experiments conducted on real-world datasets, including Douban, Amazon, and Industry, demonstrate that our proposed strategy markedly enhances recommendation accuracy, achieving approximately a 3.67% improvement in HR@10 and NDCG@10 metrics while maintaining strong privacy protection. These findings suggest that enhancing the effectiveness of cross-domain recommendation systems under privacy constraints is a promising direction for future research.

In today’s era of information overload, efficiently extracting valuable information from a large volume of textual data has become a crucial challenge in reading. This study aimed to explore the enhancement of Chinese readability in the digital domain through intelligent typesetting method. This study involved two experiments. The purpose of the first experiment was to achieve the machine learning-based assessment of the importance of individual words in Chinese articles and to automate typesetting based on the importance of words. For the second experiment, readability tests and eye-tracking reading tests were performed and the reading performance and reading attention between intelligent typesetting and general typesetting style was compared. This work proposed three Chinese typesetting methods that distinguished the importance of Chinese text information, based on font size and color. The results showed that first, when reading Chinese text, visual attention was more likely to be drawn to larger font sizes, darker brightness, or warmer-colored characters. Second, intelligent Chinese typography that distinguishes information importance through font size, color brightness, and color hue can enhance Chinese reading comprehension accuracy and subjective evaluation. It was concluded that using the TextRank model to distinguish importance of Chinese vocabulary and intelligent typesetting methods based on visual features of font could obviously improve text readability. Specifically, readability achieved through intelligent typesetting method distinguishing information importance through font color brightness surpasses that of general typesetting significantly. These intelligent typesetting methods can be widely applied in Chinese reading scenarios, such as web pages, e-books, information visualization, and other Chinese reading contexts.

Serverless computing has gained popularity for its ability to shift the burden of server management from developers to cloud providers, which allows providers to exercise greater control over resource management, optimizing configurations to enhance efficiency and performance. The diversity of serverless computing tasks, from short-lived, event-driven tasks to more complex workloads, highlights the growing importance of efficient file I/O performance for I/O-intensive workloads, yet effectively handling ephemeral storage for I/O-intensive tasks remains a challenge. Traditional file system approaches often introduce substantial latency and fail to fully leverage available memory resources within the execution environment, limiting performance and efficiency. Our work stems from the observation of the under-utilization of memory resources in serverless computing platforms and the potential efficiency improvement of I/O operations using an in-memory file system. Based on this observation, we propose Ephemera , a system designed to enhance ephemeral storage efficiency and memory utilization. Ephemera satisfies three design goals: transparent memory I/O integration , heterogeneous tasks resource synergy , and harmonized cluster workload orchestration . Ephemera integrates three components: the Runtime Daemon, responsible for managing a container’s in-memory file system; the Tenant Manager, facilitating memory configuration sharing across containers; and the Cluster Controller, optimizing workload balancing. Our experiments demonstrate that Ephemera significantly improves performance for I/O-intensive tasks compared to traditional file systems. Specifically, Ephemera decreases I/O processing time by 50% on average and reduces latency by up to 95.73% in certain scenarios with negligible overhead.

Non-parallel story author-style transfer is an important but challenging task in natural language process, which requires transferring an input story into another author-style while maintaining source semantics. Despite recent progress, current text style transfer systems still face the challenges of low robustness of the model and low quality of the generated stories. To address these challenges, we propose an end-to-end framework incorporating dual encoder components and a fusion mechanism, which can achieve explicit style-content disentanglement and effectively fusing source-domain content with target-domain stylistic features. First, we extract text from source stories containing content information using empirical extraction rules and prompt engineering. And then, we propose a novel generation model which achieves story-style transfer through capturing source content features and target style features and then fusing them. We use two additional training objectives to learn high-level discourse representations. Moreover, we have constructed a new dataset for this task. Extensive experiments based on automatic and human evaluation show that our model significantly outperforms state-of-the-art baselines, achieving approximately 8.5% average improvement in comprehensive performance metrics, demonstrating the effectiveness of our model in story-style transfer.

Sequential recommender systems often struggle with accurate personalized recommendations due to data sparsity issues. Existing works use variational autoencoders and generative adversarial network methods to enrich sparse data. However, they often overlook diversity in the latent data distribution, hindering the model’s generative capacity. This characteristic of generative methods can introduce additional noise in many cases. Moreover, retaining personalized user preferences through the generation process remains a challenge. This work introduces DGFedRS, a Federated Recommender System Based on Diffusion Augmentation and Guided Denoising, designed to capture the diversity in the latent data distribution while preserving user-specific information and suppressing noise. In particular, we pre-train the diffusion model using the recommender dataset and use a diffusion augmentation strategy to generate interaction sequences, expanding the sparse user-item interactions in the discrete space. To preserve user-specific preferences in the generated interactions, we employ a guided denoising strategy to guide the generation process during reverse diffusion. Subsequently, we design a noise control strategy to reduce the damage to personalized information during the diffusion process. Additionally, a stepwise scheduling strategy is devised to input generated data into the sequential recommender model based on their challenge levels. The success of the DGFedRS approach is demonstrated by thorough experiments conduct on three real-world datasets.

The Graph Convolutional Network (GCN) is a powerful method for handling graph data in deep learning, which has found extensive application and exhibited outstanding performance in recommendation systems. Graph Contrastive Learning (GCL), on the other hand, is a self-supervised learning approach that learns valuable information about graph structures by contrasting representations of diverse elements within the graph. In the realm of recommendation systems, Graph Contrastive Learning has garnered significant research attention and interest.This paper introduces a novel framework for the Contrastive Learning model, termed Deeper Graph Contrastive Learning with Attention mechanism (ADGCL). To address the issue of excessive smoothing of nodes caused by the stacking of multiple layers of graph convolutional layers, resulting in reduced differences between nodes and subsequently impacting the performance of tasks such as node classification, this paper employs crosslayer connections and representation mapping. These techniques mitigate the oversmoothing phenomenon, enabling the network to delve deeper into learning. Furthermore, the incorporation of a self-attention mechanism for features enhances the efficiency of the network in handling information between different feature graphs, thereby improving model performance. The core idea revolves around enabling the model to adaptively learn the weights of feature graphs, better capturing crucial features to enhance overall performance. Experimental results demonstrate that our model outperforms current state-of-the-art methods, achieving a performance improvement of nearly 5% on the Yelp2018, Douban-Book, and Ml-1M datasets, respectively.

Knowledge graphs (KGs) have demonstrated exceptional effectiveness within the domain of recommendation systems.However, in recommendation systems, the high-quality representation of KG is hindered by noise and data sparsity.Additionally, user-item(UI) interactions dominate item node representations, with minimal influence from the knowledge graph.Furthermore, single-level contrastive learning (CL) inadequately captures implicit information in node embeddings.In order to address those issue, we propose the Hierarchy Knowledgeaware Contrastive Learning (HKCL) framework, which utilizes view enhancement in CL.Firstly, we employ an edge learner to reduce noise interference during the learning process by refining the representation of the graph.Secondly we encode the UI interaction graph and KG using Graph Neural Network (GNN).Finally, during the CL process, we use a finer-grained hierarchical CL method to enhance node representations, thereby discovering more potential features to alleviate the issue of data sparsity.Specifically, we conduct CL at the user-item level, item level, and entity-item level, making the CL more compatible with recommendation learning.Empirical findings from three extensive real-world datasets illustrate that our approach improves the accuracy of recommendation result.

For the purpose of improving the user experience of mobile smart terminals for the elderly, accessibility tools have become an indispensable component of intelligent system frameworks to meet the varying needs of senior citizens for assistance. Nonetheless, the complexity of accessibility assistive tools complicates the lives of some elderly with a lack of cognitive experience to use these tools independently. Our purpose is to study the relationship between the accessibility design elements of smartphones, the emotions, and touch behaviors of elderly users, and to optimize the accessibility design methods of mobile smart terminals. First, we obtained the four interface elements with the most influence on users through a Semantic Differential evaluation involving 50 participants. Second, in a touch event collection experiment, 20 participants’ touch data was collected, and an SVM-based correlation model between interface elements, user emotions, and touch behaviors was established. Third, the prototype of the assistive recommender was iterated employing the combination of the correlation model and Genetic Algorithms. Ultimately, the effectiveness of the six touch media recommendations was compared through usability testing. The experimental results indicates that: (1) Emotion-aware accessibility assistive recommendation systems based on haptic recognition can enhance the elderly’s ability to access information through mobile terminals. (2) Middle-aged and elderly users have a more robust negative emotion reflection for dragging and swiping on touch screens. (3) Negative emotions of elderly users can assist identify design defects of accessibility tools. Our research work distilled a set of design suggestions for digital accessibility improvement, thereby enhancing the usability and inclusiveness of assistive tools in diverse contexts, and reducing the psychological pressure caused by unfamiliar interfaces. This study supplies a reference for improving the emotional experience of elderly people employing mobile terminals and extending the approach to accessible design.

Whole Slide Images (WSIs) are paramount in the medical field, with extensive applications in disease diagnosis and treatment. Recently, many deep-learning methods have been used to classify WSIs. However, these methods are inadequate for accurately analyzing WSIs as they treat regions in WSIs as isolated entities and ignore contextual information. To address this challenge, we propose a novel Dual-Granularity Cooperative Diffusion Model (DCDiff) for the precise classification of WSIs. Specifically, we first design a cooperative forward and reverse diffusion strategy, utilizing fine-granularity and coarse-granularity to regulate each diffusion step and gradually improve context awareness. To exchange information between granularities, we propose a coupled U-Net for dual-granularity denoising, which efficiently integrates dual-granularity consistency information using the designed Fine- and Coarse-granularity Cooperative Aware (FCCA) model. Ultimately, the cooperative diffusion features extracted by DCDiff can achieve cross-sample perception from the reconstructed distribution of training samples. Experiments on three public WSI datasets show that the proposed method can achieve superior performance over state-of-the-art methods. The code is available at https://github.com/hemo0826/DCDiff.

The 23rd International Conference on Computer and Information Science (ICIS 2023) is sponsored by the Institute of Electrical and Electronics Engineers (IEEE) and the International Association for Computer and Information Science (ACIS) and in cooperation with Jiangnan University, China

Raining is a common weather condition, and may seriously degrade the performances of outdoor computer vision systems, such as surveillance and autonomous navigation. Rain streaks may exhibit diverse appearances in the captured images, depending on their distances from the camera. For example, sparse rain streaks near the camera lens may appear as continuous and translucent strips, while distant densely accumulated rain streaks are more like fog and mist. Existing rain removal methods are mainly based on a single input image. However, on a single image, it is difficult to estimate a reliable depth map for rain removal. A light field image (LFI) records abundant structural and texture information of the target scene by capturing multi-perspective sub-aperture views with a single exposure. With a LFI, it is easier to estimate the depth maps, and rain streak locations across sub-aperture views are highly correlated. We observe that rain streaks usually have different slops and/or chromaic values, compared with the background scene, along the epipolar plane images (EPIs) of an LFI. Thus, we propose to make use of 3D EPIs to detect rain streaks and restore the background. To this end, we propose a novel GAN architecture to remove rain streaks from an LFI. Our method takes as input a 3D EPI, i.e., a stacked of sub-aperture views along the same row of a rainy LFI. It first estimates the disparity maps for the 3D EPI by utilizing an auto-encoder based depth estimation sub-network. The disparity maps concatenated with the input sub-aperture views are then fed into a non-local residual block, and two branched autoencoder sub-networks are used to extract rain-streaks and recover rain-free sub-aperture views. Extensive experiments conducted on both synthetic real-world-like LFIs and real-world LFIs demonstrate the effectiveness of our method.

From the perspective of philosophy, ontology relations denote ultimate semantic relations of related knowledge concepts. Beyond doubt, it is still a very difficult problem on how to automatically depict and construct ontology relations because of its high abstractness. Some latest research attempted to realize ontology relation learning by learning abstract hierarchies or similarities among knowledge concepts. Inspired by the requirements of associative semantic cognition like in the human brain, a constructivist ontology relation learning (CORL) method is put forward in this study by borrowing the idea of the constructivist learning theory. Wherein, two following points are supposed: 1) each symbol knowledge is looked as a token of representing certain abstract pattern and 2) each pattern denotes a type of relation structures on other patterns, or a directly observed event data, such as physical sensing data, natural image, sound data, text word etc. So, ontology relation could be considered as the associative support degrees from other knowledge concepts to the target concept, which reflects how one knowledge ontology can be demarcated by other knowledge concepts. Then, the knowledge network can be employed to represent an entire domain knowledge system. Meanwhile, an associative random walk mechanism (ARWM) on knowledge network can be considered to explain the semantic generative process of every document. Thus, CORL can be realized by integrating ARWM into an extended latent Dirichlet allocation (LDA) model. Some theoretical and experimental analysis are done. The corresponding results demonstrate that CORL can obtain effective associative semantic relations among concept words, and gain some novel characteristics in better representing knowledge ontology than existing methods.

Chest X-ray (CXR) is commonly performed as an initial investigation in COVID-19, whose fast and accurate diagnosis is critical. Recently, deep learning has a great potential in detecting people who are suspected to be infected with COVID-19. However, deep learning resulting with black-box models, which often breaks down when forced to make predictions about data for which limited supervised information is available and lack inter-pretability, still is a major barrier for clinical integration. In this work, we hereby propose a semantic-powered explainable model-free few-shot learning scheme to quickly and precisely diagnose COVID-19 with higher reliability and transparency. Specifically, we design a Report Image Explanation Cell (RIEC) to exploit clinically indicators derived from radiology reports as interpretable driver to introduce prior knowledge at training. Meanwhile, multi-task collaborative diagnosis strategy (MCDS) is developed to construct N-way K-shot tasks, which adopts a cyclic and collaborative training approach for producing better generalization performance on new tasks. Extensive experiments demonstrate that the proposed scheme achieves competitive results (accuracy of 98.91%, precision of 98.95%, recall of 97.94% and F1-score of 98.57%) to diagnose COVID-19 and other pneumonia infected categories, even with only 200 paired CXR images and radiology reports for training. Furthermore, statistical results of comparative experiments show that our scheme provides an interpretable window into the COVID-19 diagnosis to improve the performance of the small sample size, the reliability and transparency of black-box deep learning models. Our source codes will be released on https://github.com/AI-medical-diagnosis-team-of-JNU/SPEMFSL-Diagnosis-COVID-19.

Computed tomography (CT) provides information for diagnosis, PET attenuation correction (AC), and radiation treatment planning (RTP). Disadvantages of CT include poor soft tissue contrast and exposure to ionizing radiation. While MRI can overcome these disadvantages, it lacks the photon absorption information needed for PET AC and RTP. Thus, an intelligent transformation from MR to CT, i.e., the MR-based synthetic CT generation, is of great interest as it would support PET/MR AC and MR-only RTP. Using an MR pulse sequence that combines ultra-short echo time (UTE) and modified Dixon (mDixon), we propose a novel method for synthetic CT generation jointly leveraging prior knowledge as well as partial supervision (SCT-PK-PS for short) on large-field-of-view images that span abdomen and pelvis. Two key machine learning techniques, i.e., the knowledge-leveraged transfer fuzzy c-means (KL-TFCM) and the Laplacian support vector machine (LapSVM), are used in SCT-PK-PS. The significance of our effort is threefold: 1) Using the prior knowledge-referenced KL-TFCM clustering, SCT-PK-PS is able to group the feature data of MR images into five initial clusters of fat, soft tissue, air, bone, and bone marrow. Via these initial partitions, clusters needing to be refined are observed and for each of them a few additionally labeled examples are given as the partial supervision for the subsequent semi-supervised classification using LapSVM; 2) Partial supervision is usually insufficient for conventional algorithms to learn the insightful classifier. Instead, exploiting not only the given supervision but also the manifold structure embedded primarily in numerous unlabeled data, LapSVM is capable of training multiple desired tissue-recognizers; 3) Benefiting from the joint use of KL-TFCM and LapSVM, and assisted by the edge detector filter based feature extraction, the proposed SCT-PK-PS method features good recognition accuracy of tissue types, which ultimately facilitates the good transformation from MR images to CT images of the abdomen-pelvis. Applying the method on twenty subjects' feature data of UTE-mDixon MR images, the average score of the mean absolute prediction deviation (MAPD) of all subjects is 140.72 ± 30.60 HU which is statistically significantly better than the 241.36 ± 21.79 HU obtained using the all-water method, the 262.77 ± 42.22 HU obtained using the four-cluster-partitioning (FCP, i.e., external-air, internal-air, fat, and soft tissue) method, and the 197.05 ± 76.53 HU obtained via the conventional SVM method. These results demonstrate the effectiveness of our method for the intelligent transformation from MR to CT on the body section of abdomen-pelvis.

We attempt to intelligently identify violations of throwing waste from vehicles (TWV) in real-time traffic surveillance videos. In addition to polluting the environment, TWV easily causes injury to sanitation workers responsible for cleaning roads by passing vehicles. However, manual inspection is still the commonest way to recognize such uncivilized behavior in videos with very high time and labor-consuming. In answer to these challenges, we design a novel 20-layer residual network (Nov-ResNet-20) for training the vehicle-thrown-waste identification model (VTWIM). Then, incorporating Nov-ResNet-20, Selective Search, and Non-Maximum Suppression (NMS), we propose the deep-residual-network-leveraged vehicle-thrown-waste identification method (DRN-VTWI). Our method first splits one video frame into several regions matching suspected objects marked with location boxes via Selective Search. Then, in terms of the VTWIM trained by Nov-ResNet-20 our method identifies the regions containing TWV. Last, our method removes the redundant location boxes for each recognized, vehicle-thrown waste and only keeps the best one. The significance of our work is four-fold: 1) Nov-ResNet-20 has a moderate depth: 6 convolutional layers, 7 residual layers, and in total 20 weight layers. Due to the joint contribution of the residual, batch normalization, dropout, and cross-entropy loss, it is eligible to identify TWV using a small quantity of manually-annotated training samples. 2) Selective Search diversely marks all possible, suspected objects in video frames, whereas NMS keeps the best location box for each recognized vehicle-thrown waste, removing all redundancies. In this way, DRN-VTWI finds potential violations of TWV as many as possible and optimally annotates vehicle-thrown wastes in frames as well. 3) Combining the power of Nov-ResNet-20, Selective Search, and NMS, DRN-VTWI well solves the challenging, intelligent identification of vehicle-thrown wastes for real-time traffic surveillance. Experimental studies conducted on real-time traffic surveillance videos demonstrate the effectiveness as well as superiority of our efforts.

Magnetic resonance imaging (MRI) is an important medical diagnosis technique in clinical diagnosis, while the quality of MR images is always damaged by the noise which is caused in the image acquisition process. In the classic image denoising methods, how to design an excellent regularizer with the prior knowledge of image is the key to solve the denoising problem. In this work, we introduce the deep neural regularizer for the MRI denoising tasks, the deep neural regularizer is made up of neural network structure and objective function, similar to the classic regularizer, both of these two parts are designed with the prior knowledge of image. The proposed neural network has three main parts: encoder network, decoder network and skip connections, the encoder network which consists of five down-sampling blocks is enforced to deeply extract low-resolution or highly-abstract MR image features, similar to the encoder network architecture, the decoder network is made up of five up-sampling blocks and is enforced to restore high-resolution MR image features. To generate more finer image features, we also use skip connections to transmit the abstract information from encoder to decoder directly. The objective function consists of data fidelity term and image quality penalty term, specifically, to enforce the capability of data fidelity term, we add the self-designed image structural consistency calculation to data fidelity term besides only calculating the image consistency over image pixels with mean squared error. Meanwhile, to guide the network generate more clearer image and reduce noise information, with the prior knowledge of image sharpness, an image quality penalty term which calculates the MR image sharpness is also added to the objective function. Experimental results over the simulated MRI data and real clinical data demonstrate the proposed network can achieve superior performance compared with other methods in terms of peak signal to noise ratio, structure similarity index, image average gradient and image information entropy.

Currently, the emergence of edge computing provides low-latency and high-efficiency computing for the Internet of Things (IoT). However, new architectures, protocols, and security technologies of edge computing need to be verified and evaluated before use. Since network emulation based on a cloud platform has advantages in scalability and fidelity, it can provide an effective network environment for verifying and evaluating new edge computing technologies. Therefore, we propose a high-performance emulation technology supporting the routing protocol based on a cloud platform. First, we take OpenStack as a basic network environment. To improve the performance and scalability of routing emulation, we then design the routing emulation architecture according to the software-defined network (SDN) and design the cluster scheduling mechanism. Finally, the design of the Open Shortest Path First (OSPF) protocol can support communication with physical routers. Through extensive experiments, we demonstrate that this technology not only can provide a realistic OSPF protocol but also has obvious advantages in the overhead and performance of routing nodes compared with those of other network emulation technologies. Furthermore, the realization of the controller cluster improves the scalability in the emulation scale.

The space-ground integrated network (SGIN) is an important direction of future network development and is expected to play an important role in edge computing for the Internet of Things (IoT). Through integration with an SGIN, IoT applications can provide services with long-distance and wide-coverage features. However, SGINs are typical large-scale and time-varying networks for which new network technologies, protocols, and applications must be rigorously evaluated and validated. Therefore, a reliable experimental platform is necessary for SGINs. This paper presents a cloud-based experimental platform for the SGIN context named SGIN-Stack. First, the architecture of SGIN-Stack, which combines the Systems Tool Kit (STK) and OpenStack, is described. Based on this architecture, a seamless linkage between OpenStack and STK is achieved to realize synchronous, dynamic, and real-time network emulation for an SGIN, and the dynamic differential compensation technology and a random number generation algorithm are applied to improve the emulation accuracy for satellite links. Finally, an emulation scenario is constructed that includes six space-based backbone nodes, sixty-six space-based access nodes, and a ground station. Based on this emulation scenario, experiments concerning the satellite link delays, bit error ratio (BER), and throughput are carried out to prove the high fidelity of our SGIN-Stack platform. Emulation experiments involving satellite orbital maneuvers and attitude adjustments show that SGIN-Stack can be used for dynamic and real-time SGIN emulation.

Jigsaw puzzle algorithm is important as it can be applied to many areas such as biology, image editing, archaeology and incomplete crime-scene reconstruction. But, still, some problems exist in the process of practical application, for example, when there are a large number of similar objects in the puzzle fragments, the error rate will reach 30%–50%. When some fragments are missing, most algorithms fail to restore the images accurately. When the number of fragments of the jigsaw puzzle is large, efficiency is reduced. During the intelligent puzzle, mainly the Sum of Squared Distance Scoring (SSD), Mahalanobis Gradient Compatibility (MGC) and other metrics are used to calculate the similarity between the fragments. On the basis of these two measures, we put forward some new methods: 1. MGC is one of the most effective measures, but using MGC to reassemble the puzzle can cause an error image every 30 or 50 times, so we combine the Jaccard and MGC metric measure to compute the similarity between the image fragments, and reassemble the puzzle with a greedy algorithm. This algorithm not only reduces the error rate, but can also maintain a high accuracy in the case of a large number of fragments of similar objects. 2. For the lack of fragmentation and low efficiency, this paper uses a new method of SSD measurement and mark matrix, it is general in the sense that it can handle puzzles of unknown size, with fragments of unknown orientation, and even puzzles with missing fragments. The algorithm does not require any preset conditions and is more practical in real life. Finally, experiments show that the algorithm proposed in this paper improves not only the accuracy but also the efficiency of the operation.

To provide immersive freehand painting experience, we proposed a flexible airbrush model making use of the hands tracking capability of Leap Motion Controller. The airbrush model uses a common screen as the painting canvas. When the user moves hands over the screen, the brush model continually acquires his/her hands movement data and extracts multiple control signals which describes multiple gestures. The virtual airbrush moves along with the user's hands movement as if it is fixed on his/her finger, and its properties change with gestures' change. When the virtual airbrush intersects with the screen, it continually exerts paints onto the screen. User test shows that the user can easily create multifarious brush stroke effects by directly operating over the screen.

This paper presents a new symbol-matrix-based jigsaw-puzzle algorithm for image reconstruction. The proposed algorithm first calculates the compatibility metric using the SSD (Sum of Squared Distance Scoring) between adjacent pieces. Then the algorithm constructs a matrix to express the location relationship of pieces followed by constructing a symbol matrix to record the number and rotations of pieces. Finally, we use a greed algorithm to reconstruct the images. The proposed algorithm does not require any preset conditions and can reconstruct the images rapidly. The experimental results have shown that the proposed algorithm can accurately reconstruct the images with 28% speed-up in execution time. The results also show that it's very effective to reconstruct the puzzles with missing pieces, which is a useful feature for applications such as artifact reconstruction, biological information reconstruction and incomplete crime-scene reconstruction.

This paper presents a non-photorealistic rendering method combining fuzzy color models with mosaic rendering to emulate the fuzziness of color usage. The method first triangulates the source image based on its local details, to emulate artists' methods of observing and analyzing the image structures. Then, it converts the color from the source image to obtain the fuzzy color for every triangle. Finally, it renders the triangles by a customizable fuzzy coloring strategy. The results show that the proposed method achieves good simulation of different artists' coloring strategies. Using the rendering application based on the proposed method, the user can quickly achieve many renderings in different coloring strategies by our mosaic rendering prototype.

Empowered by today's rich tools for media generation and distribution, and the convenient Internet access , streaming crowdsourced multimedia content (crowdsourced streaming, in brief) generalizes the single-source streaming paradigm by including massive contributors for a video/data channel. It calls a joint optimization along the path from crowdsourcers , through streaming servers, to the end-users to minimize the overall latency. The dynamics of the video sources, together with the globalized request demands and the high computation demand from each sourcer, make crowdsourced live streaming challenging even with powerful support from modern cloud computing. In this paper, we present a generic framework that facilitates a cost-effective cloud service for crowdsourced live streaming. Through adaptively leasing, the cloud servers can be provisioned in a fine granularity to accommodate geo-distributed video crowdsourcers. We present an optimal solution to deal with service migration among cloud instances of diverse lease prices. It also addresses the location impact to the streaming quality. To understand the performance of the proposed strategies in the real world, we have built a prototype system running over the planetlab and the Amazon/Microsoft Cloud. Our extensive experiments demonstrate that the effectiveness of our solution in terms of deployment cost and streaming quality.

This paper proposes an improved steganalysis algorithm to detect the secret message hidden in the compressed video. As majority of video steganographic algorithms modify motion vectors (MV) in inter-frame encoding to hide data, aliasing effect may be caused in the distribution of the difference between MVs in two adjacent macroblocks. This phenomenon has been observed in detecting the secret data that were added to the MVs in cover video. To exploit the correlations between the neighboring MVs so as to detect the hidden data more efficiently, we consider the joint distribution of the MV differences between one macroblock and the other two macroblocks neighboring to it. The calculated joint probability mass functions are used to distinguish the stego videos from the non-stego ones. The experimental results show that significant improvement in detection accuracy can be made by using the joint distribution of MV differences instead of the statistics calculated from two neighboring MVs as features.