With the widespread of Internet-of-Things (IoT) environment, a big data concept has emerged to handle large number generated by IoT devices. Moreover, since data-driven approaches now become important for business, markets have emerged, and are exploited major stakeholders, such as brokers service providers. Since many services applications utilize analytic methods with collected from devices, conflict issues between privacy exploitation raised, mainly categorized protection valuation...

With the widespread use of Internet Things, data-driven services take lead both online and off-line businesses. Especially, personal data draw heavy attention service providers because usefulness in value-added services. emerging big-data technology, a broker appears, which exploits sells about individuals to other third parties. Due little transparency between brokers/consumers, people think that current ecosystem is not trustworthy, new regulations with strengthening rights were...

With the evolutionary development of Internet Things (IoT), demand for delay-sensitive applications has increased. In this manner, mobile-edge computing (MEC) emerged as a promising technology to run these on mobile devices. Several studies have been conducted pricing schemes in MEC environment, and they focused amount offloaded data pricing. However, under previously proposed policy, server resource usage users is not properly reflected payment, so only try occupy resources much possible...

With the emergence of various Internet Things (IoT) technologies, energy-saving schemes for IoT devices have been rapidly developed. To enhance energy efficiency in crowded environments with multiple overlapping cells, selection access points (APs) should consider conservation by reducing unnecessary packet transmission activities caused collisions. Therefore, this paper, we present a novel energy-efficient AP scheme using reinforcement learning to address problem unbalanced load that arises...

nomaly detection, or outlier refers to identifying rare abnormal instances patterns within a dataset that deviate significantly from the expected normal behaviour. Various methods have been proposed, but most assume their training datasets take full, complete integrity. However, innocent integrity of data is not easy maintain in reality. Existing anomaly detection generally see given as single class and learn features can represent it well, this approach very vulnerable contamination. This...

The key challenge of unsupervised vehicle re-identification (Re-ID) is learning discriminative features from unlabelled images. Numerous methods using domain adaptation have achieved outstanding performance, but those still need a labelled dataset as source domain. This paper addresses an Re-ID method, which no any types dataset, through Self-supervised Metric Learning (SSML) based on feature dictionary. Our method initially extracts images and stores them in Thereafter, the dictionary,...

Detecting unusual behaviors of insiders on enterprise resource planning (ERP) systems is one the essential parts to reduce risks threatening and abusing resources by insiders. Many approaches detect based rule-based stochastic processes are currently limited empirical monitoring using manually established algorithms or probabilistic boundaries. Those need prior knowledge such as user permission guideline process data characteristics. Unfortunately, obtaining hard in practice, these not...

In this paper, to balance power supplement from the solar energy's intermittent and unpredictable generation, we design a energy generation trading platform (EggBlock) using Internet of Things (IoT) systems blockchain technique. Without centralized broker, proposed EggBlock can promote between users equipped with panels, demand generation. By applying second price sealed-bid auction, which is one suitable pricing mechanisms in technique, it possible derive truthful bidding market...

Personal data have become the key to data-driven services and applications whereas privacy requirements are now strongly imposed by regulations. Meanwhile, people find it difficult understand whether handle personal comply with their agreements Therefore, need for indicators, which summarize contents as forms of scoring, labels, etc., has increased empower users' rights providing understandable information about privacy. For firm proper criteria methods evaluating level risks compliance...

With the widespread of electric vehicles (EV), necessity installing EV charging facility has rapidly increased. Since operating are costly and complex, choosing a proper power operation scheme is critical in implementing cost-effective service. Therefore, this article proposes an efficient for by jointly analyzing not only monetary issues with various rate policies battery wear-out costs energy storage system (ESS) but also uncertainty caused users' behavior under different requirements. By...

In this paper, we propose Normality-Calibrated Autoencoder (NCAE), which can boost anomaly detection performance on the contaminated datasets without any prior information or explicit abnormal samples in training phase. The NCAE adversarially generates high confident normal from a latent space having low entropy and leverages them to predict dataset. is trained minimise reconstruction errors uncontaminated maximise samples. experimental results demonstrate that our method outperforms...

Recently, vehicle re-identification methods based on deep learning constitute remarkable achievement. However, this achievement requires large-scale and well-annotated datasets. In constructing the dataset, assigning globally available identities (Ids) to vehicles captured from a great number of cameras is labour-intensive, because it needs consider their subtle appearance differences or viewpoint variations. paper, we propose camera-tracklet-aware contrastive (CTACL) using multi-camera...

Many Internet of Things applications provide methods to create mashup services with if-then-else approaches due its simplicity. When a number are increased, it may be suffered by service conflict, that is, perform opposite actions or abnormal behavior at the same time. To overcome this issue, paper proposes concept context descriptor detect conflict and visualize their situations.

The performance of road defect segmentation (a.k.a. pixel-level detection) has been improved alongside with remarkable achievement deep learning. Those improvements need a large-scale and well-constructed dataset. However, surface materials or designs vary from country to country, the patterns defects are hard pre-define. In this paper, we propose novel multi-source domain adaptation method boost on an unlabelled proposed generates ensembled labels using transferred information models...

Reidentification (Re-id) of vehicles in a multicamera system is an essential process for traffic control automation. Previously, there have been efforts to reidentify based on shots images with identity (id) labels, where the model training relies quality and quantity labels. However, labeling vehicle ids labor-intensive procedure. Instead relying expensive we propose exploit camera tracklet that are automatically obtainable during Re-id dataset construction. In this article, present weakly...

Web-based multi-screen digital signage system is designed to show a multimedia content via multiple screens which are connected by network. In this demonstration, novel method control proposed. The proposed uses standard web technologies display same contents and seamless moving each devices screen.

Recently, as the number of smart devices increases, way digital broadcast contents is changed. This change leads that conventional media acceptsWeb platform and its services to provide more quality contents. Based on this change, in education field, broadcasting also follows trend. The traditional platforms, which just delivered a lecture one-way, are utilized Web technology make interaction between teacher student. however, insufficient satisfy users' demands for two-way interactions. paper...

The number of wireless access points (APs) have increased rapidly in recent years to support increasing demands local area network. However, some the research findings show that many these APs are actually idle mainly during off-peak hours, i.e., there is no active user within their coverage areas time. Nevertheless, identify availability users, all remained powered-on always regardless presence or absence any inside areas. To save energy hours APs, sleep mechanism could be applied. In this...

We demonstrate sigAlbum, a novel photo album service. It utilizes tiny text intelligence to intelligently classify and search relevant photos inside smartphone without any image processing. sigAlbum has three key features. First, it organizes in-device into well-defined categories. Secondly, retrieves semantically related photos. Finally, works in stand-alone, privacy-protecting manner. To the best of our knowledge, this is first work that provides intelligent classification semantic...

Personal data is becoming increasingly valuable in business, as the insights that can be obtained from processing continue to improve. However, it also cause adverse effects on individuals. To improve quality while satisfying privacy compliance, companies now have focused collecting informed consent individuals directly handle personal without applying any privacy-preserving techniques. Even though obtain use data, transparency and accountability ensure deal with according consent, necessary...

With the proliferation of Internet Things technologies, health care services that target a household equipped with IoT devices are widely emerging. In meantime, number global single households is expected to rapidly grow. Contactless radar-based sensors recently investigated as convenient and practical means collect biometric data subjects in households. this paper, collected by contactless installed elderly under uncontrolled environments analyzed, deep learning-based classification model...

For micro grids with renewable energy sources, the main goal is to optimize usage in a particular area based on prediction of consumption and production. However, error cannot be evaded it causes various problems operate microgrid system. To solve these problems, we are going propose two-stage operation model local In addition, by applying cooperative game theory Shapley-value algorithm, revenue payment determined real-time period actual contribution individual prosumer. Numerical anaylsis...