-
FISTNet: FusIon of STyle-path generative Networks for Facial Style Transfer
Authors:
Sunder Ali Khowaja,
Lewis Nkenyereye,
Ghulam Mujtaba,
Ik Hyun Lee,
Giancarlo Fortino,
Kapal Dev
Abstract:
With the surge in emerging technologies such as Metaverse, spatial computing, and generative AI, the application of facial style transfer has gained a lot of interest from researchers as well as startups enthusiasts alike. StyleGAN methods have paved the way for transfer-learning strategies that could reduce the dependency on the huge volume of data that is available for the training process. Howe…
▽ More
With the surge in emerging technologies such as Metaverse, spatial computing, and generative AI, the application of facial style transfer has gained a lot of interest from researchers as well as startups enthusiasts alike. StyleGAN methods have paved the way for transfer-learning strategies that could reduce the dependency on the huge volume of data that is available for the training process. However, StyleGAN methods have the tendency of overfitting that results in the introduction of artifacts in the facial images. Studies, such as DualStyleGAN, proposed the use of multipath networks but they require the networks to be trained for a specific style rather than generating a fusion of facial styles at once. In this paper, we propose a FusIon of STyles (FIST) network for facial images that leverages pre-trained multipath style transfer networks to eliminate the problem associated with lack of huge data volume in the training phase along with the fusion of multiple styles at the output. We leverage pre-trained styleGAN networks with an external style pass that use residual modulation block instead of a transform coding block. The method also preserves facial structure, identity, and details via the gated map** unit introduced in this study. The aforementioned components enable us to train the network with very limited amount of data while generating high-quality stylized images. Our training process adapts curriculum learning strategy to perform efficient, flexible style and model fusion in the generative space. We perform extensive experiments to show the superiority of FISTNet in comparison to existing state-of-the-art methods.
△ Less
Submitted 2 April, 2024; v1 submitted 18 July, 2023;
originally announced July 2023.
-
ChatGPT Needs SPADE (Sustainability, PrivAcy, Digital divide, and Ethics) Evaluation: A Review
Authors:
Sunder Ali Khowaja,
Parus Khuwaja,
Kapal Dev,
Weizheng Wang,
Lewis Nkenyereye
Abstract:
ChatGPT is another large language model (LLM) vastly available for the consumers on their devices but due to its performance and ability to converse effectively, it has gained a huge popularity amongst research as well as industrial community. Recently, many studies have been published to show the effectiveness, efficiency, integration, and sentiments of chatGPT and other LLMs. In contrast, this s…
▽ More
ChatGPT is another large language model (LLM) vastly available for the consumers on their devices but due to its performance and ability to converse effectively, it has gained a huge popularity amongst research as well as industrial community. Recently, many studies have been published to show the effectiveness, efficiency, integration, and sentiments of chatGPT and other LLMs. In contrast, this study focuses on the important aspects that are mostly overlooked, i.e. sustainability, privacy, digital divide, and ethics and suggests that not only chatGPT but every subsequent entry in the category of conversational bots should undergo Sustainability, PrivAcy, Digital divide, and Ethics (SPADE) evaluation. This paper discusses in detail the issues and concerns raised over chatGPT in line with aforementioned characteristics. We also discuss the recent EU AI Act briefly in accordance with the SPADE evaluation. We support our hypothesis by some preliminary data collection and visualizations along with hypothesized facts. We also suggest mitigations and recommendations for each of the concerns. Furthermore, we also suggest some policies and recommendations for EU AI policy act concerning ethics, digital divide, and sustainability.
△ Less
Submitted 27 March, 2024; v1 submitted 13 April, 2023;
originally announced May 2023.
-
Fog based Secure Framework for Personal Health Records Systems
Authors:
Lewis Nkenyereye,
S. M. Riazul Islam,
Mahmud Hossain,
M. Abdullah-Al-Wadud,
Atif Alamri
Abstract:
The rapid development of personal health records (PHR) systems enables an individual to collect, create, store and share his PHR to authorized entities. Health care systems within the smart city environment require a patient to share his PRH data with a multitude of institutions' repositories located in the cloud. The cloud computing paradigm cannot meet such a massive transformative healthcare sy…
▽ More
The rapid development of personal health records (PHR) systems enables an individual to collect, create, store and share his PHR to authorized entities. Health care systems within the smart city environment require a patient to share his PRH data with a multitude of institutions' repositories located in the cloud. The cloud computing paradigm cannot meet such a massive transformative healthcare systems due to drawbacks including network latency, scalability and bandwidth. Fog computing relieves the burden of conventional cloud computing by availing intermediate fog nodes between the end users and the remote servers. Aiming at a massive demand of PHR data within a ubiquitous smart city, we propose a secure and fog assisted framework for PHR systems to address security, access control and privacy concerns. Built under a fog-based architecture, the proposed framework makes use of efficient key exchange protocol coupled with ciphertext attribute based encryption (CP-ABE) to guarantee confidentiality and fine-grained access control within the system respectively. We also make use of digital signature combined with CP-ABE to ensure the system authentication and users privacy. We provide the analysis of the proposed framework in terms of security and performance.
△ Less
Submitted 12 November, 2020;
originally announced November 2020.
-
Blockchain-Enabled EHR Framework for Internet of Medical Things
Authors:
Lewis Nkenyereye,
S. M. Riazul Islam,
Mahmud Hossain,
M. Abdullah-Al-Wadud,
Atif Alamri
Abstract:
The Internet of Medical Things (IoMT) offers an infrastructure made of smart medical equipment and software applications for health services. Through the internet, the IoMT is capable of providing remote medical diagnosis and timely health services. The patients can use their smart devices to create, store and share their electronic health records (EHR) with a variety of medical personnel includin…
▽ More
The Internet of Medical Things (IoMT) offers an infrastructure made of smart medical equipment and software applications for health services. Through the internet, the IoMT is capable of providing remote medical diagnosis and timely health services. The patients can use their smart devices to create, store and share their electronic health records (EHR) with a variety of medical personnel including medical doctors and nurses. However, unless the underlying combination within IoMT is secured, malicious users can intercept, modify and even delete the sensitive EHR data of patients. Patients also lose full control of their EHR since most health services within IoMT are constructed under a centralized platform outsourced in the cloud. Therefore, it is appealing to design a decentralized, auditable and secure EHR system that guarantees absolute access control for the patients while ensuring privacy and security. Using the features of blockchain including decentralization, auditability and immutability, we propose a secure EHR framework which is mainly maintained by the medical centers. In this framework, the patients' EHR data are encrypted and stored in the servers of medical institutions while the corresponding hash values are kept on the blockchain. We make use of security primitives to offer authentication, integrity and confidentiality of EHR data while access control and immutability is guaranteed by the blockchain technology. The security analysis and performance evaluation of the proposed framework confirms its efficiency.
△ Less
Submitted 11 November, 2020;
originally announced November 2020.
-
IoT service slicing and task offloading for edge computing
Authors:
JaeYoung Hwang,
Lionel Nkenyereye,
NakMyoung Sung,
JaeHo Kim,
JaeSeung Song
Abstract:
With the advancement of IoT technology, various domains such as smart factories, smart cities and smart cars use the IoT to provide value-added services. In addition, technologies such as MEC and network slicing provide another opportunity for the IoT to support more advanced and real-time services that could not have been previously supported. However, the simple integration of such technologies…
▽ More
With the advancement of IoT technology, various domains such as smart factories, smart cities and smart cars use the IoT to provide value-added services. In addition, technologies such as MEC and network slicing provide another opportunity for the IoT to support more advanced and real-time services that could not have been previously supported. However, the simple integration of such technologies into the IoT does not take full advantage of MEC and network slicing or the reduction of latency and traffic prioritization, respectively. Therefore, there is a strong need for an efficient integration mechanism for IoT platforms to maximize the benefit of using such technologies. In this article, we introduce a novel architectural framework that enables the virtualization of an IoT platform with minimum functions to support specific IoT services and host the instance in an edge node, close to the end-user. As the instance provides its service at the edge node where the MEC node and network slice are located, the traffic for the end-user does not need to traverse back to the cloud. This architecture guarantees not only low latency but also efficient management of IoT services at the edge node. To show the feasibility of the proposed architecture, we conduct an experimental evaluation by comparing the transmission time of both IoT services running on the central cloud and those using sliced IoT functions in the edge gateway. The results show that the proposed architecture provides two times faster transmission time than that from the conventional cloud-based IoT platform.
△ Less
Submitted 24 August, 2020;
originally announced August 2020.
-
Amateur Drones Detection: A machine learning approach utilizing the acoustic signals in the presence of strong interference
Authors:
Zahoor Uddin,
Muhammad Altaf,
Muhammad Bilal,
Lewis Nkenyereye,
Ali Kashif Bashir
Abstract:
Owing to small size, sensing capabilities and autonomous nature, the Unmanned Air Vehicles (UAVs) have enormous applications in various areas, e.g., remote sensing, navigation, archaeology, journalism, environmental science, and agriculture. However, the unmonitored deployment of UAVs called the amateur drones (AmDr) can lead to serious security threats and risk to human life and infrastructure. T…
▽ More
Owing to small size, sensing capabilities and autonomous nature, the Unmanned Air Vehicles (UAVs) have enormous applications in various areas, e.g., remote sensing, navigation, archaeology, journalism, environmental science, and agriculture. However, the unmonitored deployment of UAVs called the amateur drones (AmDr) can lead to serious security threats and risk to human life and infrastructure. Therefore, timely detection of the AmDr is essential for the protection and security of sensitive organizations, human life and other vital infrastructure. AmDrs can be detected using different techniques based on sound, video, thermal, and radio frequencies. However, the performance of these techniques is limited in sever atmospheric conditions. In this paper, we propose an efficient unsupervise machine learning approach of independent component analysis (ICA) to detect various acoustic signals i.e., sounds of bird, airplanes, thunderstorm, rain, wind and the UAVs in practical scenario. After unmixing the signals, the features like Mel Frequency Cepstral Coefficients (MFCC), the power spectral density (PSD) and the Root Mean Square Value (RMS) of the PSD are extracted by using ICA. The PSD and the RMS of PSD signals are extracted by first passing the signals from octave band filter banks. Based on the above features the signals are classified using Support Vector Machines (SVM) and K Nearest Neighbor (KNN) to detect the presence or absence of AmDr. Unique feature of the proposed technique is the detection of a single or multiple AmDrs at a time in the presence of multiple acoustic interfering signals. The proposed technique is verified through extensive simulations and it is observed that the RMS values of PSD with KNN performs better than the MFCC with KNN and SVM.
△ Less
Submitted 28 February, 2020;
originally announced March 2020.
-
RDSP: Rapidly Deployable Wireless Ad Hoc System for Post-Disaster Management
Authors:
Ajmal Khan,
Adnan Munir,
Zeeshan Kaleem,
Farman Ullah,
Muhammad Bilal,
Lewis Nkenyereye,
Shahen Shah,
Long D. Nguyen,
S. M. Riazul Islam,
Kyung-Sup Kwak
Abstract:
In post-disaster scenarios, such as after floods, earthquakes, and in war zones, the cellular communication infrastructure may be destroyed or seriously disrupted. In such emergency scenarios, it becomes very important for first aid responders to communicate with other rescue teams in order to provide feedback to both the central office and the disaster survivors. To address this issue, rapidly de…
▽ More
In post-disaster scenarios, such as after floods, earthquakes, and in war zones, the cellular communication infrastructure may be destroyed or seriously disrupted. In such emergency scenarios, it becomes very important for first aid responders to communicate with other rescue teams in order to provide feedback to both the central office and the disaster survivors. To address this issue, rapidly deployable systems are required to re-establish connectivity and assist users and first responders in the region of incident. In this work, we describe the design, implementation, and evaluation of a rapidly deployable system for first response applications in post-disaster situations, named RDSP. The proposed system helps early rescue responders and victims by sharing their location information to remotely located servers by utilizing a novel routing scheme. This novel routing scheme consists of the Dynamic ID Assignment (DIA) algorithm and the Minimum Maximum Neighbor (MMN) algorithm. The DIA algorithm is used by relay devices to dynamically select their IDs on the basis of all the available IDs of networks. Whereas, the MMN algorithm is used by the client and relay devices to dynamically select their next neighbor relays for the transmission of messages. The RDSP contains three devices; the client device sends the victim's location information to the server, the relay device relays information between client and server device, the server device receives messages from the client device to alert the rescue team. We deployed and evaluated our system in the outdoor environment of the university campus. The experimental results show that the RDSP system reduces the message delivery delay and improves the message delivery ratio with lower communication overhead.
△ Less
Submitted 20 January, 2020;
originally announced January 2020.
-
Software-Defined Network-Based Vehicular Networks: A Position Paper on Their Modeling and Implementation
Authors:
Lionel Nkenyereye,
Lewis Nkenyereye,
S M Riazul Islam,
Yoon Ho Choi,
Muhammad Bilal,
Jong Wook Jan
Abstract:
There is a strong devotion in the automotive industry to be part of a wider progression towards the Fifth Generation (5G) era. In-vehicle integration costs between cellular and vehicle-to-vehicle networks using Dedicated Short Range Communication could be avoided by adopting Cellular Vehicle-to-Everything (C-V2X) technology with the possibility to re-use the existing mobile network infrastructure.…
▽ More
There is a strong devotion in the automotive industry to be part of a wider progression towards the Fifth Generation (5G) era. In-vehicle integration costs between cellular and vehicle-to-vehicle networks using Dedicated Short Range Communication could be avoided by adopting Cellular Vehicle-to-Everything (C-V2X) technology with the possibility to re-use the existing mobile network infrastructure. More and more, with the emergence of Software Defined Networks, the flexibility and the programmability of the network have not only impacted the design of new vehicular network architectures but also the implementation of V2X services in future intelligent transportation systems. In this paper, we define the concepts that help evaluate software-defined-based vehicular network systems in the literature based on their modeling and implementation schemes. We first overview the current studies available in the literature on C-V2X technology in support of V2X applications. We then present the different architectures and their underlying system models for LTE-V2X communications. We later describe the key ideas of software-defined networks and their concepts for V2X services. Lastly, we provide a comparative analysis of existing SDN-based vehicular network system grouped according to their modeling and simulation concepts. We provide a discussion and highlight vehicular ad-hoc networks' challenges handled by SDN-based vehicular networks.
△ Less
Submitted 31 August, 2019;
originally announced September 2019.
-
A Fuzzy System based Approach to Extend Network Lifetime for En-Route Filtering Schemes in WSNs
Authors:
M. K. Shahzad,
Lewis Nkenyereye,
S. M. Riazul Islam
Abstract:
Wireless sensor networks suffer from false report injection attacks. This results in energy drain over sensor nodes on the event traversal route. Novel en-route filtering schemes counter this problem by filtering these attacks on designated verification nodes. However, these filtering schemes among other limitations inherently are network lifetime inefficient. Generally, report traversal paths and…
▽ More
Wireless sensor networks suffer from false report injection attacks. This results in energy drain over sensor nodes on the event traversal route. Novel en-route filtering schemes counter this problem by filtering these attacks on designated verification nodes. However, these filtering schemes among other limitations inherently are network lifetime inefficient. Generally, report traversal paths and verification nodes are also fixed. In this paper, we cater these limitations in our proposed scheme. Simulation experiments results show that proposed schemes outperforms existing en-route filtering schemes in networks lifetime. We employed a Fuzzy Logic System to select forwarding nodes from candidate nodes based on current network conditions. Proposed scheme gains in network lifetime, and energy-efficiency while having comparable false report filtering efficiency.
△ Less
Submitted 20 December, 2018;
originally announced December 2018.