Feed-forward 3D Gaussian splatting (3DGS) models have gained significant popularity due to their ability to generate scenes immediately without needing per-scene optimization. Although omnidirectional images are becoming more popular since they reduce the computation required for image stitching to composite a holistic scene, existing feed-forward models are only designed for perspective images. The unique optical properties of omnidirectional images make it difficult for feature encoders to correctly understand the context of the image and make the Gaussian non-uniform in space, which hinders the image quality synthesized from novel views. We propose OmniSplat, a training-free fast feed-forward 3DGS generation framework for omnidirectional images. We adopt a Yin-Yang grid and decompose images based on it to reduce the domain gap between omnidirectional and perspective images. The Yin-Yang grid can use the existing CNN structure as it is, but its quasi-uniform characteristic allows the decomposed image to be similar to a perspective image, so it can exploit the strong prior knowledge of the learned feed-forward network. OmniSplat demonstrates higher reconstruction accuracy than existing feed-forward networks trained on perspective images. The code is available on: https://github.com/esw0116/OmniSplat.

This paper proposes a novel fire detection and suppression system (FDSS) designed to detect and extinguish fires in the nacelle of a wind turbine. The FDSS incorporates three sensors: an infrared camera, an optical camera, and a 3D LiDAR, as well as a fire suppression system mounted on a pan and tilt control system. The FDSS features three key characteristics. First, an ensemble learning network simultaneously classifies and detects fire/smoke regions by integrating a classification neural network, an object detection neural network, and a cumulative alarm. This novel architecture significantly improves fire detection accuracy and reduces false alarm rates. Second, multimodal information precisely localizes overheat and fire/smoke regions, enabling the FDSS to automatically aim and extinguish fires by controlling the pan and tilt system. Third, a graph‐based neural network accurately classifies the affected components in the nacelle using point cloud data from the 3D LiDAR. This novel neural network for object classification provides sufficient information for the location of a fire accident. Field and virtual experiments conducted in a fire test room and virtual nacelle environments demonstrate the FDSS’s effectiveness. Quantitative comparisons of three deep learning networks further highlight that these neural networks outperform other state‐of‐the‐art deep learning models. Consequently, the FDSS provides a cost‐effective and autonomous surveillance solution, enhancing the safe operation of wind turbines with advanced technologies in the fourth industrial revolution.

We propose an efficient framework for 3D human pose and shape estimation from a video, named Uncertainty-Guided SpatioTemporal Transformer (UNSPAT). Unlike previous video-based methods that consider temporal relationships with global average pooled features, our approach incorporates both spatial and temporal dimensions without compromising spatial information. We address the excessive complexity of spatiotemporal attention through two modules: Spatial Alignment Module (SAM) and Space2Batch. The modules align input features and compute temporal attention at every spatial position in a batch-wise manner. Furthermore, our uncertainty-guided attention re-weighting module improves performance by diminishing the impact of artifacts. We demonstrate the effectiveness of the UNSPAT on widely used benchmark datasets and achieve state-of-the-art performance. Our method is robust to challenging scenes, such as occlusion, and cluttered backgrounds, showing its potential for real-world applications.

Single-image 3D human reconstruction aims to reconstruct the 3D textured surface of the human body given a single image. While implicit function-based methods recently achieved reasonable reconstruction performance, they still bear limitations showing degraded quality in both surface geometry and texture from an unobserved view. In response, to generate a realistic textured surface, we propose ReFu, a coarse-to-fine approach that refines the projected backside view image and fuses the refined image to predict the final human body. To suppress the diffused occupancy that causes noise in projection images and reconstructed meshes, we propose to train occupancy probability by simultaneously utilizing 2D and 3D supervisions with occupancy-based volume rendering. We also introduce a refinement architecture that generates detail-preserving backside-view images with front-to-back warping. Extensive experiments demonstrate that our method achieves state-of-the-art performance in 3D human reconstruction from a single image, showing enhanced geometry and texture quality from an unobserved view.

The instrumentation and control (I&C) system of a nuclear power plant (NPP) employs a cybersecurity program regulated by the government. Through regulation, the government requires the implementation of security controls in order for a system to be developed and operated. Accordingly, the licensee of an NPP works to comply with this requirement, beginning in the development phase. The compliance-driven approach is efficient when the government supervises NPPs, but it is inefficient when a licensee constructs them. The security controls described in regulatory guidance do not consider system characteristics. In other words, the development organization spends a considerable amount of time excluding unnecessary control items and preparing the evidence to justify their exclusion. In addition, security systems can vary according to the developer’s level of security knowledge, leading to differences in levels of security between systems. This paper proposes a method for a developer to select the appropriate security controls when preparing the security requirements during the early development phase; it is designed to ensure the system’s security and reduce the cost of excluding unnecessary security controls. We have formalized the representation of attack patterns and security control patterns and identified the relationships between these patterns. We conducted a case study applying RG 5.71 in the Plant Protection System (PPS) to confirm the validity of the proposed method.

The task of image-based virtual try-on aims to transfer a target clothing item onto the corresponding region of a person, which is commonly tackled by fitting the item to the desired body part and fusing the warped item with the person. While an increasing number of studies have been conducted, the resolution of synthesized images is still limited to low (e.g., 256×192), which acts as the critical limitation against satisfying online consumers. We argue that the limitation stems from several challenges: as the resolution increases, the artifacts in the misaligned areas between the warped clothes and the desired clothing regions become noticeable in the final results; the architectures used in existing methods have low performance in generating high-quality body parts and maintaining the texture sharpness of the clothes. To address the challenges, we propose a novel virtual try-on method called VITON-HD that successfully synthesizes 1024×768 virtual try-on images. Specifically, we first prepare the segmentation map to guide our virtual try-on synthesis, and then roughly fit the target clothing item to a given person’s body. Next, we propose ALIgnment-Aware Segment (ALIAS) normalization and ALIAS generator to handle the misaligned areas and preserve the details of 1024×768 inputs. Through rigorous comparison with existing methods, we demonstrate that VITON-HD highly surpasses the baselines in terms of synthesized image quality both qualitatively and quantitatively.

An outlier is an object that is considerably dissimilar with the remainder of the dataset. In this paper, we first propose the notion of centrality and center-proximity as novel outlierness measures which can be considered to represent the characteristics of all of the objects in the dataset. We then propose a graph-based outlier detection method which can solve the problems of local density, micro-cluster, and fringe objects. Finally, through extensive experiments, we show the effectiveness of the proposed method.

A variety of recommendation methods have been proposed to satisfy the performance and accuracy; however, it is fairly difficult to satisfy both of them because there is a trade-off between them. In this paper, we introduce the notion of category experts and propose the recommendation method by exploiting the ratings of category experts instead of those of the users similar to a target user. We also extend the method that uses both the category preference of a target user and his/her similarity to category experts. We show that our method significantly outperforms the existing methods in terms of performance and accuracy through extensive experiments with real-world data.

This paper proposes a Computational Intelligence framework for network security management. The framework uses reinforcement learning and Bayesian methods to achieve cross-layer optimization in heterogeneous, multi-layer wireless and wireline networks. Metrics based on the reputation of a node are used to measure performance. OPNET simulations results indicate that routing algorithms based on the newly proposed framework exhibit nearly four times improved network performance measured in terms of goodput.

Recent advances in the ubiquitous computing and the context-aware system have driven the intelligent agents for the management of user movements and multi contexts. With the growth of the agent-based system and the advances in networking technologies, the design and development of the multi agent system have risen in importance. However, the conventional models offer some kinds of static services to limited users and operate the sequential structure of multi agent middleware. Relatively few studies have been devoted to the research issues of the convergence in heterogeneous networks, the diversity in service requirements and the efficiency in service management to effectively support next-generation context-aware services. Therefore, we propose a parallel Multi Agent Middleware which provides efficient approaches to manage the seamless service and the reconstruction of services according to user movements and a couple of situations. Our architecture allows us to create a rich set of MsMu (Multi Source Multi Use) services through the extension of OsMu (One Source Multi Use) services for enhancing a service satisfaction. Moreover, we enhance the aspect of the service response time and the success probability of the context-aware schedule by using the determination rules of service reconstruction.

In smart home, people who use of context-aware systems should expect to be more dynamic and adaptive services by analyzing residents along with life styles. Currently, the research effort is focused on the service-oriented structure and ontology-based context model. However, the conventional model did not consider the classified service, service reconstruction and definition of service type by classified service patterns. In this paper, we propose a session manager which manages service registration states and classified service according to the user movements. Moreover, a session manager offers to analyze the service pattern by changing environments and user location. Our system could create the dynamic services according to service pattern analysis. Therefore, our system reduces the offered service delays and improves the maximum 64% performance of service response time compared with non-existence the session manager. We show the feasibility of our system by evaluating the efficiency on a test bed.

Traditional hop-by-hop dynamic routing makes inefficient use of network resources as it forwards packets along already congested shortest paths while uncongested longer paths may be underutilized. To maintain network-wide load balancing, we propose Autonomous Network management with Team learning based Self-configuration (ANTS) which attempts to manage a feasible route for traffic flow with QoS constraints in heterogeneous networks. To enable cognitive intelligence for network-wide load balancing, we implement a cross-layer mechanism in which learning agents in middleware layer can monitor the queue sizes of MAC layer, thereby allowing for the discovery of optimal routes. We present OPNET simulation results illustrating that, in comparison to original OSPF and AODV (2.18 Mbits/s with 46.46% packet loss rate), ANTS dramatically achieves a higher packet delivery (9.57 Mbits/s with 0.53% packet loss rate).

This paper presents the goal, accomplishments and research issues of an NSF project. The project aims to develop a distributed event-triggered knowledge sharing network (ETKnet) for government organizations to share, not only data and application operations, but also knowledge embedded in organizational and inter-organizational policies, regulations, data and security constraints as well as collaborative processes and operating procedures. A unified knowledge and process specification language has been developed to formally specify multi-faceted human and organizational knowledge in terms of three types of knowledge rules and rule structures. A user-friendly interface is provided for collaborating organizations to define events of interest as well as application operations, knowledge rules, rule structures and triggers. Events are published in a global registry for browsing, querying, event subscription and notification. Rules and rule structures are automatically translated into Web services for distributed processing in ETKnet. Event data are dots that can be connected dynamically across organizational boundaries through the interoperation of knowledge rules, processes and application operations.

This paper presents a framework to address the combination of ubiquitous computing interests as an intelligent assisting location-aware service and development of its hardware and software systems. In our real testbed, the system can dynamically localize the areas of interest and adapt to changes in the ubiquitous intelligence landscape.

This paper proposes a context model which separates the upper and lower layer according to characteristic of each class using the Web Ontology Language (OWL). We also present a context-aware multi-agent system (CAMAS) architecture which provides the solution of policy problem using the classification of rule and implementation of a rule-based engine. We enhance the response time for offering adaptive services. We solve the policy conflict problem among users by using reasoning and privilege delegation.

The paper presents an information infrastructure for supporting Internet-based scalable e-business enterprises (ISEE). The information infrastructure is formed by a network of ISEE hubs, each of which has a number of replicable e-business servers providing various e-services to individuals and businesses. The servers are the implementations of a number of core technologies developed to facilitate collaborative e-business, including business event and rule management, active distributed objects, constraint satisfaction processing, and cost benefit analysis and selection. Using the e-services provided by these servers, other e-services such as constraint-based brokering, supplier selection, active business process management, and automated negotiation can be developed. Supply chain management is used as an example of collaborative e-business in the descriptions of these technologies and their implementations.

A data warehouse stores information that is collected from multiple, heterogeneous information sources for the purpose of complex querying and analysis. Information in the warehouse is typically stored in the form of materialized views, which represent pre-computed portions of frequently asked queries. One of the most important tasks when designing a warehouse is the selection of materialized views to be maintained in the warehouse. The goal is to select a set of views in such a way as to minimize the total query response time over all queries, given a limited amount of time for maintaining the views (maintenance-cost view selection problem). In this paper, we propose an efficient solution to the maintenance-cost view selection problem using a genetic algorithm for computing a near-optimal set of views. Specifically, we explore the maintenance-cost view selection problem in the context of OR view graphs. We show that our approach represents a dramatic improvement in time complexity over existing search-based approaches using heuristics. Our analysis shows that the algorithm consistently yields a solution that lies within 10% of the optimal query benefit while at the same time exhibiting only a linear increase in execution time. We have implemented a prototype version of our algorithm which is used to simulate the measurements used in the analysis of our approach.

The Internet and distributed object technologies have made it possible for different business enterprises to draw upon the best of their resources for conducting joint business as a virtual e-enterprise (VEE). To enable virtual e-enterprises, the integration of legacy applications and the modeling and enactment of concurrent business processes are necessary. The authors combine the features of the messaging approach and the distributed object approach to system integration. Business Object Documents (BOD) are used for transmitting business operations and data among application systems. Message transmission is supported by two underlying communication infrastructures: CORBA and Java RMI. The separation of messaging from communication infrastructure allows the underlying infrastructure to be changed without impacting application systems. Also, business processes are modeled as sequences or network structures of BOD transmissions. The process models are replicated at all sites and used by an extended information infrastructure to enable distributed, concurrent enactment of processes.

With the emergence of e-business as the next killer application for the Web, automating bargaining-type negotiations between clients (i.e., buyers and sellers) has become increasingly important. With IDEAL (Internet based Dealmaker for e-business), we have developed an architecture and framework, including a negotiation protocol, for automated negotiations among multiple IDEAL servers. The main components of IDEAL are a constraint satisfaction processor (CSP) to evaluate a proposal, an Event-Trigger-Rule (ETR) server for managing and triggering the execution of rules which make up the negotiation strategy (rules can be updated at run-time to deal with the dynamic nature of negotiations), and a cost-benefit analysis to help in the selection of alternative strategies. We have implemented a fully functional prototype system of IDEAL to demonstrate automated negotiations among buyers and suppliers participating in a supply chain.