International Journal of Computer (IJC)

Intelligent Clock Gating for FPGA-based RISC Architectures: A Novel Approach to Switching Activity and Dynamic Power Reduction

Prasanth Varasala — 2024-07-10

In modern digital systems, dynamic power consumption remains a critical concern, particularly in Field-Programmable Gate Arrays (FPGAs) utilized in power-sensitive applications. This paper presents a novel intelligent clock gating technique specifically tailored for FPGA-based RISC architectures to effectively reduce switching activity and dynamic power dissipation. Our approach leverages a combination of hardware and software strategies to dynamically control the clock signals to inactive modules, thereby minimizing unnecessary power consumption. The proposed method integrates seamlessly with existing FPGA design flows and RISC architectures, providing a scalable and efficient solution for power management. Through comprehensive simulations and experimental evaluations on standard benchmark circuits, we demonstrate a significant reduction in dynamic power consumption while maintaining performance and functionality. At higher frequencies overall 64% power on total power is saved.

Multi-Class Cancer Classification with SVM Using Wrapper Forward and Backward Feature Selection for Dimension Reduction

May Myat Myat Khaing — 2024-07-01

The use of machine learning (ML) into healthcare has shown enormous growth in recent years. The efficacy of supervised ML models is significantly influenced by the quality of the training data. Feature selection is a crucial factor that affects the performance of machine learning models, especially in complex tasks like multi-class cancer classification. This research investigates the efficacy of using forward feature selection and backward feature elimination approaches in combination with logistic regression. The features generated using these approaches are then used for cancer type classification using support vector machines (SVM).The focus of our study is to use a partially complete gene dataset obtained from the Indian Council of Medical study (ICMR) for the purpose of classifying different types of cancer using Support Vector Machines (SVM). Our approach demonstrated a remarkable success rate of 96% when using features selected via the forward selection method and 97% when using features obtained through the backward selection method in multi-class cancer classification.

Comparative Analysis of Machine Learning Algorithms for Diabetes Prediction: Finding the Optimal Approach

Aftab UL Nabie — 2024-06-21

Diabetes, as a chronic disease, poses a rapidly escalating risk to human health, stemming from a complex interplay of factors such as obesity, elevated blood glucose levels, and various other triggers. Central to its onset is the disruption of insulin hormone function, resulting in abnormal metabolism and increased blood sugar levels. In this paper, we propose a solution to this pressing issue using machine learning techniques. By applying various machine learning algorithms on the Pima Indian diabetes (PID) dataset, we aim to identify the most effective algorithm for this task. Leveraging powerful machine learning algorithms such as (SVM) Support Vector Machine, (RF) Random Forest and others, we endeavor to forecast the onset of diabetes. Through the amalgamation of these techniques, our objective is to proactively identify individuals at risk, enabling timely intervention and preventive measures to safeguard health. The primary goal of this initiative is to mitigate the risk of diabetes onset by forecasting individuals' susceptibility and advocating for lifestyle and dietary adjustments. This study has dual objectives: firstly, to develop and implement a predictive model for diabetes using machine learning techniques, and secondly, to explore effective strategies for achieving success in this endeavor.

Assessing the Effect of Gamification in Increasing the Mastery Level of Grade 8 Students in Technology and Livelihood Education

Marissa Shirley Rose Castro — 2024-07-13

This study aims to investigate whether the gamification in lesson can help on increasing the mastery level of grade 8 students. Gamification, it refers to the application game design elements to an educational setting where competition, points, badges, and rewards helps enhance students’ engagement and motivation. The main goal is to make learning more engaging and interesting specially for learners with short attention span. The research involved a pre-test at the beginning of the quarter, discussion of the lessons with the use of gamification and post-test at the end of the quarter to collect data if the used of gamification really helps in increasing the mastery level of students. Hence, the data supports the assertion that the use of gamification in TLE 8 lessons has resulted in a significant enhancement in student performance. The significant difference between pre-test and post-test scores indicates that gamification has positively impacted the students' learning outcomes.

Comparison of Single-Shot and Two-Shot Deep Neural Network Models for Pose Estimation in Assistance Living Application

May Phyo Ko — 2024-07-01

Estimating human posture from an image or video is an essential task in computer vision. This task has detected body key points from a camera for body posture and gesture recognition technology, which enables the following applications: assisted living in the case of fall detection, yoga pose identification, character animation, and an autonomous drone control system. The rapid development of AI-based posture estimation algorithms for picture recognition has resulted in the availability of quick and dependable solutions for recognizing the human body joint in collected films. One major issue in human posture assessment is the system’s capacity to perform with high accuracy in real-time under shifting ambient conditions. The ultimate goal of the proposed transfer learning-based posture estimation assignment is to achieve real-time speed with virtually no drop accuracy. In this research paper, assisted living program (ALP) is implemented by using a single-shot deep estimation network and a pose key points angular feature. Experimental results show that transfer learning-based pose identifies and estimates posture with a frame rate of about 30 frames per second and a detection accuracy rate of 96.81%.

Optimisation of University Examination Timetable Using Hybridised Genetic and Greedy Algorithms: A Case Study of Computer Science Department, University of Ibadan

Sunday J. Agbolade — 2024-06-15

Timetable scheduling is an important aspect of decision-making in any organisation, particularly in academia. An examination timetable is expected to coordinate students, invigilators, courses, examination hall allocation, and time slots. However, the problem could be viewed as a Nondeterministic Polynomial (NP); NP-hard problem, scheduling problem has plagued humanity since its inception. Due to the complex structure of the problem in terms of hard and soft-constraints, most organisations schedule time inefficiently using manual approach. This study introduced an algorithms hybridisation method of genetic and greedy algorithms to automate the timetable scheduling process efficiently. A genetic algorithm is a heuristic search technique based on Charles Darwin's theory of natural evolution. The fitness of each course, venue, and faculty content is determined by the probabilistic optimisation which is the solution candidate in the initial population of all the objects. Subsequently, the greedy algorithm's activities selector selects the best solution. The output demonstrates that the method effectively handled all the constraints associated with timetable scheduling. Hybridising the two algorithms to build a scheduling system, such as the examination timetable. Therefore, it is a viable option to combine genetic and greedy algorithms to have an optimised examination timetable that is flexible to any situation.

Realistic Sketch-based Face Photo Synthesis using Generative Adversarial Networks (GANs)

Hnin Ei Ei Cho — 2024-06-15

Facial photo-image synthesis and sketch-based face recognition are highly advantageous, particularly in the fields of security forces and forensics. Furthermore, it makes it more feasible for law enforcement to reduce the number of possible suspects in criminal identification operations. However, since pencil drawings and photographs have different properties by nature, creating a synthesis of photographs based on sketches presents a difficult topic. In the last few decades, generative adversarial network-based systems have achieved enormous advances towards improving the performance of image synthesis. It can speed up identification times while improving matching outcomes by reducing gaps among sketch and photo representations. We perform investigations on the well-known photo-sketch pair database CUHK. First, we demonstrate how a generative adversarial network transforms hand-drawn sketches into realistic photos. Secondly, we employ suspect identification by using the pre-trained VGG16-based feature extractor network and KNN classifier. Our technique focuses on the use of deep learning-based networks, which are well-known for their capacity to process data and extract hierarchical features. The presented image-to-image translation framework minimizes the modality differences between hand-drawn face sketches and color images while improving visual quality. Tests on sketch-photo matching demonstrate significant improvements over current state-of-the-art methods on the challenging task of matching sketches with corresponding photos.