As the global transition towards renewable energy sources accelerates, solar power becomes an increasingly important solution. Identifying and understanding the current distribution of solar panel installations is crucial for future planning and decision-making process. This paper introduces SolarDetector, a transformer-based neural network model, which we developed and fine-tuned for the accurate detection of solar panels. It achieves 91.0% mIoU for the task of masking solar panels on SWISSIMAGE dataset.

Maps provide various sources of information. An important example of such information is textual labels such as cities, neighborhoods, and street names. Although we treat this information as facts, and despite the massive effort done by providers to continuously improve their accuracy, this data is far from perfect. Discrepancies in textual labels rendered on the map are one of the major sources of inconsistencies across map providers. These discrepancies can have significant impacts on the reliability of the derived information and decision-making processes. Thus, it is important to validate the accuracy and consistency in such data. Most providers treat this data as their propriety data and it is not available to the public, thus we cannot compare the data directly. To address these challenges, we introduce a novel computer vision-based approach for automatically extracting and classifying labels based on the visual characteristics of the label, which indicates its category based on the format convention used by the specific map provider. Based on the extracted data, we detect the degree of discrepancies across map providers. We consider three map providers: Bing Maps, Google Maps, and OpenStreetMaps. The neural network we develop classifies the text labels with an accuracy up to 93% in all providers. We leverage our system to analyze randomly selected regions in different markets. The studied markets are USA, Germany, France, and Brazil. Experimental results and statistical analysis reveal the amount of discrepancies across map providers per region. We calculate the Jaccard distance between the extracted text sets for each pair of map providers, which represents the discrepancy percentage. Discrepancies percentages as high as 90% were found in some markets.

We demonstrate MapsVision, a computer vision-based framework capable of identifying discrepancies across different map providers for similar geographical locations. In this study, we primarily focus on three map providers including: (a) Bing Maps, (b) Google Maps, and (c) OpenStreetMap. MapsVision detects textual data discrepancies such as: (1) missing location labels (2) misspelled or different keywords, (3) shifted labels, and (4) level of significance manifested by text or label font-size and color. For a given location, our MapsVision framework compares textual labels based on a ground truth entered manually to those that exist in the three map providers. We then use the results of the textual extraction to determine the accuracy of textual data appearing on map providers. Our framework intelligently identifies the set of techniques for each map providers' that can maximize the overall detection accuracy. MapsVision is composed of three main building blocks including: (a) a capturing module that captures map tiles from map providers, (b) an analysis tool that uses computer vision and text-analytic techniques, and (c) a rich visualization interface for displaying statistical and real-time analytics. The objective of MapsVision is to help map editors improve the textual quality of their maps compared to other map providers.

Today, people's lives are enriched by the integration of electronic maps via smartphones. Electronic maps are required for a variety of commercial activities, such as catering, movie viewing, and tourism. Route planning and navigation are particularly intrinsically linked to electronic maps. As a result, it is critical that the roads on the electronic map are complete and accurate. At the present time, there are discrepancies between the map roads of various providers. This paper evaluates the roads on various map providers' maps. Due to the varied terrain depicted on the map, assessing the road properties can be challenging. Additionally, roads of varying thicknesses exist within a tile image, making it difficult to quantify the map's road lengths. This paper proposes a method for extracting road segments using an image binarization technique and employs edge erosion to assist in automatically computing the length of roads within maps. Throughout the paper, we provide comparison and statistical analysis on using our proposed road length detection model across map providers. Results show that our detection model can identify road length accurately and hence provide an overall measure of quality of maps.

As more computing operations shift from the cloud to edge environments, the need for reliable and efficient resource allocation becomes inevitable. Unlike the cloud, edge computing environments often are equipped with limited computational capabilities which makes the task allocation process time consuming and challenging. When allocating resources, it is imperative to consider multi-criteria based on a number of factors including task requirements and the availability of existing edge-based computational capabilities. To this extent, we consider the resource allocation process across edge environments as an optimization problem that can be solved using multi-criteria decision analysis methods (MCDA). In this paper, we present an extension to our Edgify dynamic resource provisioning model that incorporates the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) for enhancing the decision making when provisioning edge-based resources across distributed edge or fog environments. We evaluate our proposed Edgify solution through multiple experiments which demonstrate the effectiveness of our proposed decision-making optimization approach.

Collecting data from smart Internet of Things (IoT) devices is becoming an increasingly essential part of many of the existing industrial applications. The importance of this data collection relies on the fact that it can uncover valuable insights and enable smarter, faster decision making. This enables organizations to quickly adapt to changes in the workflows, reduce downtime and expand the production capacity and enhance the overall operating efficiency. The problem, however, is that many of these Industrial IoT (IIoT) applications can considerably be influenced by the composition of RESTFul APIs and the microservices architecture they integrate. In addition, IIoT applications do not take into consideration the dynamism of service-based environments and requirements. To overcome these challenges, it is essential to consider the Quality of Service (QoS) characteristics of RESTful APIs and microservices particularly that these properties may fluctuate during their lifecycle. In this paper, we introduce a quality-aware microservices' architecture that continuously monitors the behavior of these services in delivering the required functionality. This paper presents experimental validation results and analysis of the presented ideas.

The detecting heartbeat abnormalities (i.e. arrhythmia) depends mainly on the examination of ECG signals over an adequate sampling period. This sampling period needs to contain sufficient data that can be extracted as features. Such features provide accurate measures for the diagnosis of heart arrhythmias. The problem, however, is that the analysis of ECG data requires to properly detect arrhythmias requires many ECG samples to be collected from patients and requires the extract of many features (e.g. temporal or morphological properties). In this paper, we introduce a neural network based solution that can detect heartbeat abnormalities with aim to minimize the feature-set required during the analysis process. Throughout the paper, we present results from testing our neural network using the MIT/BIH Arrhythmia database which show an accuracy rate of 98.70% success rate. We also provide insights to efficiently classify heartbeat rhythms as normal, bradycardia or tachycardia.

Addressing environmentally safe management of waste is becoming increasingly a challenging task. The predicament of the rate at which waste is generated due to increasing populations is also contributing to this challenge. One possible approach for effectively handling waste can be achieved by source reduction and recycling. The problem, however, improving the collection of waste can be costly particularly during the source separation process after waste is collected. It would be desirable if there exists a mechanism that can help municipalities, local governments or waste management companies to monitor in real-time sources of violations prior to the waste collection process. In this paper, we introduce recycle.io, an Internet of Things (IoT)-enabled waste management system that is based on a serverless architecture that can identify these sources of violations. Using recycle.io, it is then possible to track the violations geographically which can help local governments, for example, to improve or enforce tighter regulations for waste disposal. Our recycle.io system uses Microsoft Azure IoT Hub for device management. Throughout the paper, we demonstrate usefulness of using our approach for urban waste management in smart cities.

Identifying cardiac abnormalities has mainly been determined by the observation of electrocardiogram (ECG) signals. To collect ECG signals, it is often necessary to place ECG electrodes on the body for critical analysis of ECG data transmitted by such electrodes. By analyzing this collected data, it is then possible, for example, to examine the intervals between the heartbeats (or R-R intervals) to measure the heart rate variability (HRV). However, this process requires a multilayered setup for both hardware and software which can be costly and time consuming. To overcome these challenges, we introduce in this paper a real-time millimeter-wave radar-based, non-contact vital sign monitoring system that is capable of detecting the heart variability rate without the use of any heart rate sensors or wires required. Through this system, it is then possible to detect any heart rate abnormalities by analyzing the collected data. Throughout the paper, we present results for three individuals and compare our approach to heart rate monitoring devices and Apple Watch.

As the number of cloud services continues to increase, selecting services of interest across one or more cloud service environments using existing service selection methods raises a number of concerns such as performance, efficiency, end-to-end reliability and most importantly quality of search results. Clients often spend a considerable amount of time manually reading cloud providers' documentation to determine services that can meet their objectives and satisfy the application's requirements. Furthermore, cloud service providers' Quality of Service (QoS) claims for published services might not always be trustworthy and current cloud service selection methods do not take into consideration the dynamism of cloud environments as they are constantly changing. In addressing these challenges, we developed the Cloud Application Management (CAM), a multilayered framework that employs a meta-heuristic approach that is based on QoS for cloud services (QSCS) for enabling clients to effectively manage and control the quality of their applications deployed in the cloud. CAM supports the self-adaptive nature of the service selection process and adapts to the changes in clients' requirements and interests.

In recent years, more emphasis on how to predict traffic of web pages has increased significantly and prompted the need for exploring various methods on how to effectively forecast future values of multiple times series. In this paper, we apply a forecasting model for the purpose of predicting web traffic. In particular, we use existing Web Traffic Time Series Forecasting dataset by Google to predict future traffic of Wikipedia articles. Predicting web traffic can help web site owners in many ways including: (a) determining an effective strategy for load balancing of web pages residing in the cloud, (b) forecasting future trends based on historical data and (c) understanding the user behavior. To achieve the goals of this research work, we built a time-series model that utilizes RNN seq2seq model. We then investigate the use of symmetric mean absolute percentage error (SMAPE) for measuring the overall performance and accuracy of the developed model. Finally, we compare the outcome of our developed model to existing ones to determine the effectiveness of our proposed method in predicting future traffic of Wikipedia articles.

Searching for Web service access points is no longer attached to service registries as Web search engines have become a new major source for discovering Web services. In this work, we conduct a thorough analytical investigation on the plurality of Web service interfaces that exist on the Web today. Using our Web Service Crawler Engine (WSCE), we collect metadata service information on retrieved interfaces through accessible UBRs, service portals and search engines. We use this data to determine Web service statistics and distribution based on object sizes, types of technologies employed, and the number of functioning services. This statistical data can be used to help determine the current status of Web services. We determine an intriguing result that 63% of the available Web services on the Web are considered to be active. We further use our findings to provide insights on improving the service retrieval process.

As Web services proliferate, size and magnitude of UDDI Business Registries (UBRs) are likely to increase. The ability to discover Web services of interest then across multiple UBRs becomes a major challenge specially when using primitive search methods provided by existing UDDI APIs. Clients do not have the time to endlessly search accessible UBRs for finding appropriate services particularly when operating via mobile devices. Finding services of interest should be time effective and highly productive. This paper addresses issues relating to the efficient access and discovery of Web services across multiple UBRs and introduces a novel exploration engine, the Web Service Crawler Engine (WSCE). WSCE is capable of crawling multiple UBRs, and enables for the establishment of a centralized Web services repository that can be used for discovering Web services much more efficiently. The paper presents experimental validation, results, and analysis of the proposed ideas.

Major research challenges in discovering Web services include, provisioning of services across multiple or heterogeneous registries, differentiating between services that share similar functionalities, improving end-to-end Quality of Service (QoS), and enabling clients to customize the discovery process. Proliferation and interoperability of this multitude of Web services have lead to the emergence of new standards on how services can be published, discovered, or used (i.e. UDDI, WSDL, SOAP). Such standards can potentially provide many of these features and much more, however, there are technical challenges associated with existing standards. One of these challenges is the client.s ability to control the discovery process across accessible service registries for finding services of interest. This work proposes a solution to this problem and introduces the Web Service Relevancy Function (WsRF) used for measuring the relevancy ranking of a particular Web service based on QoS metrics and client preferences. We present experimental validation, results, and analysis of the presented ideas.