Journal of Information Systems and Telecommunication
Volume:12 Issue: 4, Oct-Dec 2024

The expansion of touch-screen devices has provided the possibility of human-machine interactions in the form of free-hand drawings. In sketch-based image retrieval (SBIR) systems, the query image is a simple binary design that represents the mental image of a person with the rough shape of an object. A simple sketch is convenient and efficient for recording ideas visually, and can outdo hundreds of words. The objective is to retrieve a natural image with the same label as the query sketch. This article presents a multi-step training method. Regression functions are used in the deep network structure to improve system performance, and various loss functions are employed for a better convergence of the retrieval system. The convolutional neural network used has two branches, one related to the sketch and the other related to the image, and these two branches can have the same or different architecture. After four training steps, a 56.48% MAP was achieved, indicating the desirable performance of the network.

One of the biometric detection methods is to identify people based on speech signals. The implementation of a speaker identification (SI) system can be done in many different ways, and recently, many researchers have been focusing on using deep neural networks. One of the types of deep neural networks is recurrent neural networks, where memory and recurrent parts are handled by layers such as LSTM or Gated Recurrent Unit (GRU). In this paper, we propose a new structure as a classifier in the speaker identification system, which significantly improves the recognition rate by combining a convolutional neural network with two layers of GRU (CNN+ GRU). MFCC coefficients that have been extracted as cell arrays from each period of Pt speech will be used as sequence vectors for the input of proposed classifier.  The performance of the SI system has improved in comparison to basic methods according to experiments conducted on two databases, LibriSpeech and VoxCeleb1. When Pt is longer, the system performs better, so that on the LibriSpeech database with 251 speakers, recognition accuracy is equal to 92.94% for Pt=1s, and it rises to 99.92% for Pt=9s. The proposed CNN+GRU classifier has a low sensitivity to specific genders, which can be said to be almost zero.

Cloud computing and fog computing are deployed as computing storage and services for the end-users. Fog computing promotes task performance through storage, computing, and networking services. Instead of taking place in centralized cloud computing data centers, these services can be provided via near-edge devices. Efficient load balancing in distributed computing systems has been the main challenge. The load balancing algorithm has an important role in enhancing the Quality of Service (QoS), throughput, and resource utilization and diminishing the potential cost and its strategy and architecture completely depend on the centralized or distributed architecture of the system and the type of requests. Cloud computing and fog computing use centralized and distributed architectures, respectively. The load balancing algorithm in these two environments cannot be the same. Meanwhile, the demand for near real-time processing requests is drastically increasing; load balancing should be able to handle real-time requests. This paper reviews and investigates the modern and diverse load balancing aspects of fog and cloud computing systems. We also categorize the load balancing algorithms in cloud and fog computing: meta-heuristic algorithms, heuristic algorithms, learning algorithms, and customized algorithms. We propose different research classes about the algorithm's type, objectives, simulation tools, and so forth. This review demonstrates that the most prevalent categories of methods used in load balancing in fog and cloud computing are custom approaches and meta-heuristic algorithms, respectively. While the most renowned load balancing algorithms have not yet succeeded in fog environments, meta-heuristic algorithms have shown their competence in cloud environments impeccably.

Nowadays, reducing energy consumption in cloud computing is of great interest due to the high operational costs and its impact on climate change. The consolidation solution is an effective method for minimizing the number of physical machines (PMs) and reducing energy consumption. The virtual machine consolidation process encounters the challenge of reducing energy consumption while effectively managing resource allocation. The aim of this paper is to address these challenges through the classification of PMs and the use of the dragonfly algorithm. The quartile parameter is utilized to classify PMs into three categories: underloaded, medium load, and overloaded. First, we identified the overloaded PMs in the overloaded category. Then, we presented a solution to select virtual machines from an overloaded PM based on resource usage. Additionally, the Dragonfly algorithm is utilized to select destinations for hosting migrant virtual machines in the medium load category. Furthermore, we identified underloaded PMs in the underloaded categories using this algorithm. The proposed solution is evaluated using the CloudSim toolkit and tested with workloads consisting of over a thousand data points from virtual machines based on PlanetLab data. The results from the simulation experiments indicate that the proposed solution, while avoiding SLA violations and minimizing additional migrations, has significantly reduced energy consumption.

This study explores the application of the COBIT (Control Objectives for Information and Related Technologies) IT governance framework to enhance the ICT Regulatory Tracker (ICTRT) scores, a tool developed by the International Telecommunication Union (ITU) to assess ICT regulatory bodies across countries. Given the absence of specific improvement strategies from the ITU, this research fills a critical gap by investigating how COBIT processes can be leveraged for ICT regulation improvement. Utilizing an Automatic Content Analysis (ACA) method, we identified significant relationships between 22 out of 37 COBIT processes and ICTRT indicators, with particular emphasis on APO09, APO11, and DSS02 processes. Focus group methodology employed to validate these findings and development of a continuous improvement plan tailored for Iran's ICT regulatory body. This plan integrates 13 COBIT processes from the identified set, providing a structured approach for implementation. The findings not only highlight effective COBIT processes but also offer actionable insights for regulatory bodies aiming to enhance their regulatory quality and advance towards a digital economy.

The need for blood and the blood donors are on the rise continuously. Poor communication between blood banks and hospitals results in improper management and wastage of available blood inventory and can cause life threats. Therefore, there is an urgent need for coordination between blood banks, hospitals, and blood donors. The Design of a Secure Intelligent Blood Bank Information System (SIBBIS) is a way to align blood banks and hospitals with the help of the Internet. SIBBIS is a web application through which registered hospitals can check the availability of required blood, send a request for blood to the nearest blood bank or donor that matches the blood requirements, and order blood online as requested. A blood bank can also send a request to another blood bank in case of unavailable blood. The person willing to donate blood can find the nearest blood banks using SIBBIS. The location of the blood bank can be traced using maps. The life of the hardware, software, refrigerator cleaning of refrigerator and vaccination of employees are monitored intelligently. For this system, we developed a standard process-oriented information System analysis methodology using use case, activity, system sequence, entity relationship, and class diagrams. The usability of the developed system was evaluated, and it was 93.39%.