Research Work

Research Interests

• High-Performance Computing
• Multi-core and Parallel Architectures
• Parallel Computing and Algorithms
• Parallel Programming Models (Shared, Distributed, Hybrid): OpenMP, OpenCL, SYCL, MPI
• Performance Modelling and Engineering
• Performance Visualization Tools and Parallel Simulation Frameworks
• Domain-Specific Languages
• Modern C++ (C++11/14/17): Object-Oriented Template Meta-Programming

Research Highlights

Title: ARVA: Augmented Reality-based Visual Aids for Mobility Enhancement Through Real-time Video Stream Transformations

Status: Submitted (under review)

Journal: Computers in Human Behavior

Abstract: Visual field loss (VFL) is a chronic vision defect caused by blind spots (scotoma) on the normal VF, which can significantly affect the daily activities of patients. Despite the effectiveness of existing VR/AR-based systems as vision aids, they are restricted to a single XR device and suffer from low video quality, content loss, high levels of contradiction, and inadequate mobility evaluation. To address these issues, we developed two novel vision aids using DreamWorld and MagicLeap1 AR headsets, with video processing techniques to increase the visual perception of people with moderate to severe VFL to the level of healthy people. A novel optimal video remapping function is proposed for each headset based on its specified characteristics to map the content of the captured live video to the largest intact region of the VF map, preserving its quality with minimized blurriness and distortion. As a proof-of-concept, a comprehensive empirical user study was carried out on 44 healthy subjects whose normal VF was blocked using pre-defined artificial scotomas. Statistical analysis of the data obtained from object counting and multi-tasking walking track tests before and after applying the remapping function proves the promising performance of the prototypes for visual awareness and mobility enhancement.

Keywords: Augmented Reality, Visual filed loss, Head-mounted vision aid, Remapping function, DreamWorld glasses, Magic Leap one glasses

Title: Financial Distress Prediction in the Bank Sector Using Data Mining and Machine Learning Techniques

Status: Ready for Submission

Abstract: Nowadays, bank failures are a common and sensitive issue in a developing country like Bangladesh. So, it's important to examine and predict the financial health of a bank so that it can help minimize and rectify the upcoming or present losses of banks and customers. As we know, data mining has various uses in the prediction field, but in Bangladesh, the use of data mining techniques in the banking and financial distress sectors has rarely been seen, where previous studies worked with only the Altman z-score method. So our study focuses on using different performance matrices and also compering their performance accuracy with the model. Basically, it will help to widen the use of data mining in predicting distress in the banking sector. However,  we have collected data of 8 banks of Bangladesh and found out the important features or financial ratios and on those datasets, each of these ML techniques were tested and their performance accuracy were measured. According to the result the decision tree, random forest classifier, and gradient boosting performed better compared to SVM, logistic regression, and K-nearest neighbors in predicting distress.

Title: The Impact of Social Media on Political Awareness Among University Students

Status: Submitted (under review)

Journal: Journal of Political Communication

Abstract: Social media has become a significant platform for political engagement and information dissemination. This study investigates the role of social media in enhancing political awareness among university students in Bangladesh. By conducting a survey with 500 participants from various universities, we examined the frequency of social media use, types of political content consumed, and the correlation between social media activity and political knowledge. The findings indicate that social media platforms, particularly Facebook and Twitter, play a crucial role in shaping political opinions and increasing awareness of current events. However, the study also highlights the challenges of misinformation and echo chambers. The implications of these findings suggest the need for educational programs to promote critical thinking and media literacy among students.

Keywords: Social Media, Political Awareness, University Students, Misinformation, Media Literacy

Title: Analysis of Voter Behavior in Rural and Urban Areas of Bangladesh

Status: Ready for Submission

Abstract: Understanding voter behavior is essential for democratic processes and policy-making. This research aims to analyze the differences in voter behavior between rural and urban areas in Bangladesh. Using data from the 2022 national elections, we performed a comparative analysis focusing on demographic factors, access to information, and socio-economic status. The study employed statistical methods to identify patterns and trends in voting behavior. Results revealed significant differences in voter priorities and decision-making processes, with rural voters showing a stronger inclination towards local issues and candidates, while urban voters prioritized national policies and party affiliations. The study's findings contribute to a deeper understanding of electoral dynamics and can inform strategies for political campaigns and voter education programs.

Keywords: Voter Behavior, Rural and Urban Differences, Electoral Dynamics, Bangladesh, Political Campaigns

Title: The Role of Political Campaigns in Shaping Public Opinion

Status: In Progress

Abstract: Political campaigns are critical in shaping public opinion and influencing electoral outcomes. This study explores the effectiveness of various campaign strategies used in the 2022 national elections in Bangladesh. Through content analysis of campaign materials, media coverage, and public speeches, we assessed the impact of message framing, emotional appeals, and issue salience on voter perceptions. Additionally, we conducted focus group discussions with voters to gain insights into their responses to different campaign tactics. Preliminary results suggest that campaigns focusing on positive messaging and local community engagement were more successful in garnering voter support. This research underscores the importance of strategic communication in political campaigning and its potential to sway public opinion.

Keywords: Political Campaigns, Public Opinion, Message Framing, Emotional Appeals, Electoral Outcomes.