Bilgisayar Mühendisliği Bölümü Koleksiyonu

Permanent URI for this collectionhttps://hdl.handle.net/20.500.12573/203

Browse

Now showing 1 - 20 of 191

Active Subnetwork GA: A Two Stage Genetic Algorithm Approach to Active Subnetwork Search
(BENTHAM SCIENCE PUBL LTDEXECUTIVE STE Y-2, PO BOX 7917, SAIF ZONE, 1200 BR SHARJAH, U ARAB EMIRATES, 2017) Ozisik, Ozan; Bakir-Gungor, Burcu; Diri, Banu; Sezerman, Osman Ugur; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü; Bakir-Gungor, Burcu
Background: A group of interconnected genes in a protein-protein interaction network that contains most of the disease associated genes is called an active subnetwork. Active subnetwork search is an NP-hard problem. In the last decade, simulated annealing, greedy search, color coding, genetic algorithm, and mathematical programming based methods are proposed for this problem. Method: In this study, we employed a novel genetic algorithm method for active subnetwork search problem. We used active node list chromosome representation, branch swapping crossover operator, multicombination of branches in crossover, mutation on duplicate individuals, pruning, and two stage genetic algorithm approach. The proposed method is tested on simulated datasets and Wellcome Trust Case Control Consortium rheumatoid arthritis genome-wide association study dataset. Our results are compared with the results of a simple genetic algorithm implementation and the results of the simulated annealing method that is proposed by Ideker et al. in their seminal paper. Results and Conclusion: The comparative study demonstrates that our genetic algorithm approach outperforms the simple genetic algorithm implementation in all datasets and simulated annealing in all but one datasets in terms of obtained scores, although our method is slower. Functional enrichment results show that the presented approach can successfully extract high scoring subnetworks in simulated datasets and identify significant rheumatoid arthritis associated subnetworks in the real dataset. This method can be easily used on the datasets of other complex diseases to detect disease-related active subnetworks. Our implementation is freely available at https://www.ce.yildiz.edu.tr/personal/ozanoz/file/6611/ActSubGA
Ambient Energy Harvesting for Low Powered Wireless Sensor Network based Smart Grid Applications
(IEEE, 345 E 47TH ST, NEW YORK, NY 10017 USA, 2019) Faheem, Muhammad; Ashraf, Muhammad Waqar; Butt, Rizwan Aslam; Raza, Basit; Ngadi, Md. Asri; Gungor, Vehbi Cagri; 0000-0003-4907-6359; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü
Limited battery lifetime is one of the most critical issues for wireless sensor networks (WSNs)-based smart grid (SG) applications. Recently, ambient energy harvesting (AEH) has been considered to significantly improve the network lifetime of the WSNs-based SG applications. However, extracting a significant amount of energy from the ambient energy resource due to time varying links quality affected by power grid environments is the main issue for WSNs-based applications in SG. In this paper, we propose a novel multi-source energy harvesting mechanisms for WSNs-based SG applications. The propose hybrid ambient energy harvesting framework through the designed circuitry successfully harvests massive power density by capturing the radial electric field (EF) and ambient radio frequency WiFi 2.4GHz band signals present in the vicinity of 500kV power grid station. The design energy harvesting schemes have been implemented on the recently developed routing protocol for SG applications. The experiments using EstiNet9.0, demonstrate that the designed framework is efficient in terms of energy harvesting capabilities to enable a long-lasting lifetime of the WSNs-based smart grid applications.
Analysis of Battery-Powered Sensor Node Lifetime for Smart Grid Applications
(IEEE345 E 47TH ST, NEW YORK, NY 10017 USA, 2016) Eris, Cigdem; Gungor, V. Cagri; Boluk, Pinar Sarisaray; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü; Gungor, V. Cagri
Wireless Sensor Networks (WSNs) enable smart grids where sensor nodes monitor and control the important parameters of power grid components. However, energy-aware communication protocols should be developed to extend network lifetime of WSNs in smart grid environments. In this study, the lifetime of wireless sensor nodes has been analyzed for various smart grid environments, such as 500 kV substation, main power control room, and underground network transformer vaults. In addition, the effects of different operation modes of sensor nodes on node lifetime have been reviewed.
Analysis of compressive sensing and energy harvesting for wireless multimedia sensor networks
(ELSEVIER, RADARWEG 29, 1043 NX AMSTERDAM, NETHERLANDS, 2020) Tekin, Nazli; Gungor, Vehbi Cagri; 0000-0002-4275-8544; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü
One of the main concerns of Wireless Multimedia Sensor Networks (WMSNs) is the huge data size causing the higher energy consumption in transmission. The high energy consumption is a critical problem for lifetime of network includes sensor nodes with limited battery. The data size reduction and Energy Harvesting (EH) methods are the promising solutions to improve the network lifetime. The main objective of this paper is to evaluate the impact of the different data size reduction methods, such as image compression and Compressive s Sensing (CS), and EH methods, such as vibration, thermal and indoor solar, on WMSNs lifetime in industrial environments. In addition, a novel Mixed Integer Programming (MIP) framework is proposed to maximize the network lifetime when EH, CS, and Error Control (EC) approaches are utilized together. Comparative performance results show that utilizing Binary Compressive Sensing (BCS) and Indoor Solar Harvester (ISH) extends industrial network lifetime significantly. (C) 2020 Elsevier B.V. All rights reserved.
Analyzing lifetime of energy harvesting underwater wireless sensor nodes
(WILEY, 111 RIVER ST, HOBOKEN 07030-5774, NJ USA, 2019) Erdem, H. Emre; Gungor, V. Cagri; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü
Underwater Wireless Sensor Networks (UWSNs) are utilized to monitor underwater environments that pose many challenges to researchers. One of the key complications of UWSNs is the difficulty of changing node batteries after their energy is depleted. This study aims to diminish the issues related to battery replacement by improving node lifetime. For this goal, three energy harvesting devices (turbine harvester, piezoelectric harvester, and hydrophone harvester) are analyzed to quantitate their impacts on node lifetime. In addition, two different power management schemes (schedule-driven and event-driven power management schemes) are combined with energy harvesters for further lifetime improvement. Performance evaluations via simulations show that energy harvesting methods joined by power management schemes can improve node lifetime substantially when actual conditions of Istanbul Bosporus Strait are considered. In this respect, turbine harvester makes the biggest impact and provides lifetime beyond 2000 days for most cases, while piezoelectric harvester can perform the same only for low duty cycle or event arrival values at short transmission ranges.
Analyzing lifetime of energy harvesting wireless multimedia sensor nodes in industrial environments
(ELSEVIER SCIENCE BV, PO BOX 211, 1000 AE AMSTERDAM, NETHERLANDS, 2018) Tekin, Nazli; Erdem, H. Emre; Gungor, V. Cagri; 0000-0002-4275-8544; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü
Recently, there has been a great demand for multimedia communication using Wireless Multimedia Sensor Networks (WMSNs) in industrial environments thanks to their low cost, flexibility, and rapid deployment. However, WMSNs face a major challenge of limited lifetime due to their limited battery capacity. Compared to regular data transmission, multimedia data transmission causes higher energy consumption because of larger data sizes leading to faster depletion of sensor node's batteries. The objective of this paper is to analytically quantify the impact of different energy harvesting methods based on vibration, indoor solar, and temperature difference as well as Fast-Zonal DCT and BinDCT based image compression methods on the lifetime of Telos and Mica2 sensor nodes deployed in indoor industrial environment. Performance results show that energy harvesting and image compression techniques improve lifetime of Mica2 and Telos motes by 51.8% and 25.8%, respectively when used with proper power management methods. (C) 2017 Published by Elsevier B.V.
Assessing Employee Attrition Using Classifications Algorithms
(Association for Computing Machinery, 2020) Ozdemir F.; Coskun M.; Gezer C.; Cagri Gungor V.; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü
Employees leave an organization when other organizations offer better opportunities than their current organizations. Continuity and sustenance and even completion of jobs are crucial issues for the companies not to suffer financial losses. Especially if the talented employees, who are at critical positions in the companies, leave the job, it becomes difficult for the organizations to maintain their businesses. Today, organizations would like to predict attrition of their employees and plan and prepare for it. However, the HR departments of organizations are not advanced enough to make such predictions in a handcrafted manner. For this reason, organizations are looking for new systems or methods that automatize the prediction of employee attrition utilizing data mining methods. In this study, we use IBM HR data set and apply different classification methods, such as Support Vector Machine (SVM), Random Forest, J48, LogitBoost, Multilayer Perceptron (MLP), K-Nearest Neighbors (KNN), Linear Discriminant Analysis (LDA), Naive Bayes, Bagging, AdaBoost, Logistic Regression, to predict the employee attrition. Different from exiting studies, we systematically evaluate our findings with various classification metrics, such as F-measure, Area Under Curve, accuracy, sensitivity, and specificity. We observe that data mining methods can be useful for predicting the employee attrition.
Attack-Aware Dynamic Upstream Bandwidth Assignment Scheme for Passive Optical Network
(Walter de Gruyter GmbH, 2023) Butt, Rizwan Aslam; FAHEEM, MUHAMMAD; Ashraf, M. Waqar; Khawaja, Attaullah; Raza, Basit; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü; Faheem, Muhammed
Network security is an important component of today’s networks to combat the security attacks. The passive optical network (PON) works at the medium access layer (MAC). A distributed denial of service (DDOS) attack may be launched from the network and transport layers of an Optical Network unit (ONU). Although there are various security techniques to mitigate its impact, however, these techniques cannot mitigate the impact on the MAC Layer of the PON and can cause an ONU to continuously drain too much bandwidth. This will result in reduced bandwidth availability to other ONUs and, thus, causing an increase in US delays and delay variance. In this work we argue that the impact of a DDOS attack can be mitigated by improving the Dynamic bandwidth assignment (DBA) scheme which is used in PON to manage the US bandwidth at the optical line terminal (OLT). The present DBA schemes do not have the capability to combat a security attack. Thus, this study, uses a machine learning approach to learn the ONU traffic demand patterns and presents a security aware DBA (SA-DBA) scheme that detects a rogue (attacker) ONU from its traffic demand pattern and limits its illegitimate bandwidth demand and only allows it the bandwidth assignment to it as per the agreed service level agreement (SLA). The simulation results show that the SA-DBA scheme results in up to 53%, 55% and 90% reduced US delays and up to 84%, 76% and 95% reduced US delay variance of T2, T3 and T4 traffic classes compared to existing insecure DBA schemes.
Autonomic performance prediction framework for data warehouse queries using lazy learning approach
(ELSEVIER, RADARWEG 29, 1043 NX AMSTERDAM, NETHERLANDS, 2020) Raza, Basit; Aslam, Adeel; Sher, Asma; Malik, Ahmad Kamran; Faheem, Muhammad; 0000-0001-6711-2363; 0000-0001-5569-5629; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü
Information is one of the most important assets of an organization. In recent years, the volume of data stored in organizations, varying user requirements, time constraints, and query management complexities have grown exponentially. Due to these problems, the performance modeling of queries in data warehouses (DWs) has assumed a key role in organizations. DWs make relevant information available to decision-makers; however, DW administration is becoming increasingly difficult and time-consuming. DW administrators spend too much time managing queries, which also affects data warehouse performance. To enhance the performance of overloaded data warehouses with varying queries, a prediction-based framework is required that forecasts the behavior of query performance metrics in a DW. In this study, we propose a cluster-based autonomic performance prediction framework using a case-based reasoning approach that determines the performance metrics of the data warehouse in advance by incorporating autonomic computing characteristics. This prediction is helpful for query monitoring and management. For evaluation, we used metrics for precision, recall, accuracy, and relative error rate. The proposed approach is also compared with existing lazy learning techniques. We used the standard TPC-H dataset. Experiments show that our proposed approach produce better results compared to existing techniques.
Autonomic workload performance tuning in large-scale data repositories
(SPRINGER LONDON LTD, 236 GRAYS INN RD, 6TH FLOOR, LONDON WC1X 8HL, ENGLAND, 2019) Raza, Basit; Sher, Asma; Afzal, Sana; Malik, Ahmad Kamran; Anjum, Adeel; Kumar, Yogan Jaya; Faheem, Muhammad; 0000-0002-2024-0699; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü
The workload in large-scale data repositories involves concurrent users and contains homogenous and heterogeneous data. The large volume of data, dynamic behavior and versatility of large-scale data repositories is not easy to be managed by humans. This requires computational power for managing the load of current servers. Autonomic technology can support predicting the workload type; decision support system or online transaction processing can help servers to autonomously adapt to the workloads. The intelligent system could be designed by knowing the type of workload in advance and predict the performance of workload that could autonomically adapt the changing behavior of workload. Workload management involves effectively monitoring and controlling the workflow of queries in large-scale data repositories. This work presents a taxonomy through systematic analysis of workload management in large-scale data repositories with respect to autonomic computing (AC) including database management systems and data warehouses. The state-of-the-art practices in large-scale data repositories are reviewed with respect to AC for characterization, performance prediction and adaptation of workload. Current issues are highlighted at the end with future directions.
A battery-friendly data acquisition model for vehicular speed estimation
(PERGAMON-ELSEVIER SCIENCE LTDTHE BOULEVARD, LANGFORD LANE, KIDLINGTON, OXFORD OX5 1GB, ENGLAND, 2016) Kaya, Sevgi; Kilic, Necati; Kocak, Taskin; Gungor, Cagri; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü; Gungor, Cagri
Modeling traffic flow and gathering accurate traffic congestion information are two challenging problems in smart transportation systems. Most of the traffic flow models and velocity estimation methodologies that have been proposed so far gather the data from GPS-equipped smart phones and extract the flow model based on GPS sampling. However, these approaches tend to fail in real life scenarios due to the insufficient vehicle data and unpredictable dynamics of the flow. Furthermore, utilization of GPS sensor leads to a battery drainage and hence reduces the overall system performance. In this paper, we propose a new battery-friendly data acquisition model to obtain the raw data. We then evaluate our model under various traffic conditions to determine its feasibility in vehicle speed estimation. The proposed model results in 88% location accuracy whereas it reduces the battery consumption by half. (C) 2016 Elsevier Ltd. All rights reserved.
BAUM-2: a multilingual audio-visual affective face database
(Kluwer Academic Publishers(SpringerLink), 2015) Eroglu Erdem, Cigdem; Turan, Cigdem; Aydin, Zafer; 0000-0001-7686-6298; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü; Aydin, Zafer
Access to audio-visual databases, which contain enough variety and are richly annotated is essential to assess the performance of algorithms in affective computing applications, which require emotion recognition from face and/or speech data. Most databases available today have been recorded under tightly controlled environments, are mostly acted and do not contain speech data. We first present a semi-automatic method that can extract audio-visual facial video clips from movies and TV programs in any language. The method is based on automatic detection and tracking of faces in a movie until the face is occluded or a scene cut occurs. We also created a video-based database, named as BAUM-2, which consists of annotated audio-visual facial clips in several languages. The collected clips simulate real-world conditions by containing various head poses, illumination conditions, accessories, temporary occlusions and subjects with a wide range of ages. The proposed semi-automatic affective clip extraction method can easily be used to extend the database to contain clips in other languages. We also created an image based facial expression database from the peak frames of the video clips, which is named as BAUM-2i. Baseline image and video-based facial expression recognition results using state-of-the art features and classifiers indicate that facial expression recognition under tough and close-to-natural conditions is quite challenging.
Bio-inspired routing protocol for WSN-based smart grid applications in the context of Industry 4.0
(WILEY, 111 RIVER ST, HOBOKEN 07030-5774, NJ USA, 2019) Faheem, M.; Butt, R. A.; Ashraf, M. W.; Begum, Seema; Ngadi, Md A.; Gungor, V. C.; 0000-0003-4907-6359; 0000-0003-4628-4486; 0000-0003-1591-7041; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü
Recently, the advences of Industry 4.0 have paved the way for a systematical deployment of the smart grid (SG) to manage continuously growing energy demand of the 21st century. This even allows the fourth stage of the industrial revolution in the power sector, which is known as the smart grid industry (SGI) 4.0. In SGI 4.0, the industrial wireless sensor networks (WSNs) and the Internet of Things are envisioned as key promising communication technologies for monitoring various SG applications due to their large-scale coverage, fault tolerance characteristics, and cost reduction. However, highly dynamic nature of the SG environments brings several unique challenges caused by systems and operating devices. This results in hampering the quality-of-service communication requirements for WSNs-based SG applications. In SGI 4.0, the routing infrastructure not only requires a reliable but also fulfills the communication requirements of diverse SG applications. Thus, a sophisticated, reliable and QoS-aware multi-hop communications network architecture enabling a real-time exchange of data for various WSNs-based SG applications is essential in SGI 4.0. Hence, this paper proposes a novel bio-inspired self-optimized butterfly mating optimization-based data gathering routing scheme called Self-optimized Intelligent routing protocol (SIRP) for WSNs-based SG applications. The extensive simulations reveal that the proposed scheme achieves its defined goals compared to existing routing schemes designed for WSNs-based applications.
Blockchain Based User Management System
(IEEE, 2020) Temiz, Mustafa; Soran, Ahmet; Arslan, Halil; Erel, Hilal; 0000-0003-0683-1836; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü; Soran, Ahmet
—Blokzinciri, çeşitli kriptografi teknikleri kullanılarak birbirine bağlanmış bloklar içerisinde bulunan verilerin, ağ üzerinde diğer noktalara dağıtılmasıyla oluşturulan güvenilir ve şeffaf bir yapıdır. Mevcut veri tabanı işlemlerinden farkı, yetki ve sorumlukların tek bir merkezi otoritede bulunmaması, bu yetki ve sorumlulukların ağda bulunan diğer düğümlere dağıtılarak görev paylaşımı sağlanmasıdır. Bunu sağlayabilmek için eşler arası ağ altyapısı kullanılmaktadır. Fakat bu aşamada güvenlik anlamında kimlik doğrulama işlemi temel güvenlik mekanizmalarından birini oluşturmaktadır. Bu çalışmada, blokzincirdeki hız sorunlarına çözüm olabileceği düşünülen, daha güvenilir ve güncel teknolojilerle entegre olacak şekilde çalışabilen bir kullanıcı yönetim sistemi önerilmektedir.
Blockchain-based Fog Computing Applications in Healthcare
(IEEE, 2020) Adanur, Beyhan; Bakir-Gungor, Burcu; Soran, Ahmet; 0000-0003-4983-2417; 0000-0002-2272-6270; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü; Adanur, Beyhan; Bakir-Gungor, Burcu; Soran, Ahmet
— Son zamanlarda blokzincir teknolojisinin sağlık alanında kullanımı artmıştır. Blokzincir teknolojisinin sağlık alanına getirdiği birçok yenilik olmasına rağmen, halen çözülmeyi bekleyen problemleri mevcuttur. Bu problemlere alternatif çözümler getirmesi amacıyla, sis bilişimin blokzincir teknolojisi ile birlikte kullanılması gündeme gelmiştir. Bu çalışmada, blokzincir tabanlı sis bilişim teknolojisinin sağlık alanındaki uygulamaları incelenmektedir. Sunulan çalışmanın amacı, sağlık alanında, blokzincir ve sis bilişiminin etkileşimli bir şekilde kullanımı hakkında okuyucuların fikir edinmelerini sağlamaktır. Bu amaç doğrultusunda öncelikle, sis bilişimi ve blokzincir teknolojileri tanıtılmıştır. Sonrasında, alanların birbirlerine entegrasyonu, bu teknolojilerin beraber kullanımının sağlık alanına getirdiği avantajlar ve dezavantajlar tartışılmış ve bu teknolojilerin beraber kullanımlarına dair sistem önerisinde bulunulmuştur.
BlockFaaS: Blockchain-enabled Serverless Computing Framework for AI-driven IoT Healthcare Applications
(SPRINGER, 2023) Golec, Muhammed; Gill, Sukhpal Singh; Golec, Mustafa; Xu, Minxian; Ghosh, Soumya K; Kanhere, Salil S; Uhlig, Steve; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü; Golec, Muhammed
With the development of new sensor technologies, Internet of Things (IoT)-based healthcare applications have gained momentum in recent years. However, IoT devices have limited resources, making them incapable of executing large computational operations. To solve this problem, the serverless paradigm, with its advantages such as dynamic scalability and infrastructure management, can be used to support the requirements of IoT-based applications. However, due to the heterogeneous structure of IoT, user trust must also be taken into account when providing this integration. This problem can be overcome by using a Blockchain that guarantees data immutability and ensures that any data generated by the IoT device is not modified. This paper proposes a BlockFaaS framework that supports dynamic scalability and guarantees security and privacy by integrating a serverless platform and Blockchain architecture into latency-sensitive Artificial Intelligence (AI)-based healthcare applications. To do this, we deployed the AIBLOCK framework, which guarantees data immutability in smart healthcare applications, into HealthFaaS, a serverless-based framework for heart disease risk detection. To expand this framework, we used high-performance AI models and a more efficient Blockchain module. We use the Transport Layer Security (TLS) protocol in all communication channels to ensure privacy within the framework. To validate the proposed framework, we compare its performance with the HealthFaaS and AIBLOCK frameworks. The results show that BlockFaaS outperforms HealthFaaS with an AUC of 4.79% and consumes 162.82 millijoules less energy on the Blockchain module than AIBLOCK. Additionally, the cold start latency value occurring in Google Cloud Platform, the serverless platform into which BlockFaaS is integrated, and the factors affecting this value are examined.
Building a challenging medical dataset for comparative evaluation of classifier capabilities
(ELSEVIER, 2024) Bozkurt, Berat; Coskun, Kerem; Bakal, Gokhan; 0000-0003-2897-3894; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü; Bozkurt, Berat; Coskun, Kerem; Bakal, Gokhan
Since the 2000s, digitalization has been a crucial transformation in our lives. Nevertheless, digitalization brings a bulk of unstructured textual data to be processed, including articles, clinical records, web pages, and shared social media posts. As a critical analysis, the classification task classifies the given textual entities into correct categories. Categorizing documents from different domains is straightforward since the instances are unlikely to contain similar contexts. However, document classification in a single domain is more complicated due to sharing the same context. Thus, we aim to classify medical articles about four common cancer types (Leukemia, Non-Hodgkin Lymphoma, Bladder Cancer, and Thyroid Cancer) by constructing machine learning and deep learning models. We used 383,914 medical articles about four common cancer types collected by the PubMed API. To build classification models, we split the dataset into 70% as training, 20% as testing, and 10% as validation. We built widely used machine-learning (Logistic Regression, XGBoost, CatBoost, and Random Forest Classifiers) and modern deep-learning (convolutional neural networks - CNN, long short-term memory - LSTM, and gated recurrent unit - GRU) models. We computed the average classification performances (precision, recall, F-score) to evaluate the models over ten distinct dataset splits. The best-performing deep learning model(s) yielded a superior F1 score of 98%. However, traditional machine learning models also achieved reasonably high F1 scores, 95% for the worst-performing case. Ultimately, we constructed multiple models to classify articles, which compose a hard-to-classify dataset in the medical domain.
Can artificial intelligence algorithms recognize knee arthroplasty implants from X-ray radiographs?
(MediHealth Academy, 2023) Gölgelioğlu, Fatih; Aşkın, Aydoğan; Gündoğdu, Mehmet Cihat; Uzun, Mehmet Fatih; Dedeturk, Bilge Kagan; Yalın, Mustafa; 0000-0002-8026-5003; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü; Dedetürk, Bilge Kağan
Aims: This study aimed to investigate the use of a convolutional neural network (CNN) deep learning approach to accurately identify total knee arthroplasty (TKA) implants from X-ray radiographs. Methods: This retrospective study employed a deep learning CNN system to analyze pre-revision and post-operative knee X-rays from TKA patients. We excluded cases involving unicondylar and revision knee replacements, as well as low-quality or unavailable X-ray images and those with other implants. Ten cruciate-retaining TKA replacement models were assessed from various manufacturers. The training set comprised 69% of the data, with the remaining 31% in the test set, augmented due to limited images. Evaluation metrics included accuracy and F1 score, and we developed the software in Python using the TensorFlow library for the CNN method. A computer scientist with AI expertise managed data processing and testing, calculating specificity, sensitivity, and accuracy to assess CNN performance. Results: In this study, a total of 282 AP and lateral X-rays from 141 patients were examined, encompassing 10 distinct knee prosthesis models from various manufacturers, each with varying X-ray counts. The CNN technique exhibited flawless accuracy, achieving a 100% identification rate for both the manufacturer and model of TKA across all 10 different models. Furthermore, the CNN method demonstrated exceptional specificity and sensitivity, consistently reaching 100% for each individual implant model. Conclusion: This study underscores the impressive capacity of deep learning AI algorithms to precisely identify knee arthroplasty implants from X-ray radiographs. It highlights AI’s ability to detect subtle changes imperceptible to humans, execute precise computations, and handle extensive data. The accurate recognition of knee replacement implants using AI algorithms prior to revision surgeries promises to enhance procedure efficiency and outcomes.
Capacity and spectrum-aware communication framework for wireless sensor network-based smart grid applications
(ELSEVIERRADARWEG 29, 1043 NX AMSTERDAM, NETHERLANDS, 2017) Faheem, Muhammad; Gungor, Vehbi Cagri; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü
Recently, wireless sensor networks (WSNs) have been widely recognized as a promising technology for enhancing various aspects of smart grid and realizing the vision of next-generation electric power system in a cost-effective and efficient manner. However, recent field tests show that wireless links in smart grid environments have higher packet error rates and variable link capacity because of dynamic topology changes, obstructions, electromagnetic interference, equipment noise, multipath effects, and fading. To overcome these communication challenges, in this paper, we propose a data capacity-aware channel assignment (DCA) and fish bone routing (FBR) algorithm for WSN-based smart grid applications. The proposed DCA framework deals with the channel scarcities by dynamically switching between different spectrum bands and employs a network for organizing WSN into a highly stable connected hierarchy. In addition, the proposed FBR mechanism provides robust loop free data paths and avoids high transmission cost, excessive end-to-end delay and restricts unnecessary multi-hop data transmission from the source to destination in the network. Thus, it significantly reduces the probability of data packet loss and preserves stable link qualities among sensor nodes for load balancing and prolonging the lifetime of wireless sensor networks in harsh smart grid environments. Comparative performance evaluations show that our proposed schemes outperform the existing communication architectures in terms of data packet delivery, communication delay and energy consumption.
CBI4.0: A cross-layer approach for big data gathering for active monitoring and maintenance in the manufacturing industry 4.0
(ELSEVIERRADARWEG 29, 1043 NX AMSTERDAM, NETHERLANDS, 2021) Faheem, Muhammad; Butt, Rizwan Aslam; Ali, Rashid; Raza, Basit; Ngadi, Md Asri; Gungor, Vehbi Cagri; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü; Faheem, Muhammad; Gungor, Vehbi Cagri
Industry 4.0 (I4.0) defines a new paradigm to produce high-quality products at the low cost by reacting quickly and effectively to changing demands in the highly volatile global markets. In Industry 4.0, the adoption of Internet of Things (IoT)-enabled Wireless Sensors (WSs) in the manufacturing processes, such as equipment, machining, assembly, material handling, inspection, etc., generates a huge volume of data known as Industrial Big Data (IBD). However, the reliable and efficient gathering and transmission of this big data from the source sensors to the floor inspection system for the real-time monitoring of unexpected changes in the production and quality control processes is the biggest challenge for Industrial Wireless Sensor Networks (IWSNs). This is because of the harsh nature of the indoor industrial environment that causes high noise, signal fading, multipath effects, heat and electromagnetic interference, which reduces the transmission quality and trigger errors in the IWSNs. Therefore, this paper proposes a novel cross-layer data gathering approach called CBI4.0 for active monitoring and control of manufacturing processes in the Industry 4.0. The key aim of the proposed CBI4.0 scheme is to exploit the multi-channel and multi-radio architecture of the sensor network to guarantee quality of service (QoS) requirements, such as higher data rates, throughput, and low packet loss, corrupted packets, and latency by dynamically switching between different frequency bands in the Multichannel Wireless Sensor Networks (MWSNs). By performing several simulation experiments through EstiNet 9.0 simulator, the performance of the proposed CBI4.0 scheme is compared against existing studies in the automobile Industry 4.0. The experimental outcomes show that the proposed scheme outperforms existing schemes and is suitable for effective control and monitoring of various events in the automobile Industry 4.0.

Browse

Browsing Bilgisayar Mühendisliği Bölümü Koleksiyonu by Access Right "info:eu-repo/semantics/closedAccess"