Our Customer Right Now

Title of Climate and Hydrological Features (May ‘13)

International Conference on Informatics, Electronics & Vision Read on IEEE Xplore

Research Abstract

This study analyzes temperature and rainfall data from 34 meteorological stations across Bangladesh over a 40-year period (1971–2010). The research evaluates climate change magnitude statistically and spatially.

Methodology

• Trend Analysis: Linear regression and coefficient of variation.
• Spatial Mapping: Inverse distance weighted interpolation using GIS.
• Simulation: Autoregressive integrated moving average (ARIMA) time series models.

Key Findings

• Confirmed a 0.20°C per decade upward trend in mean temperature.
• Identified erratic rainfall patterns during pre-monsoon and post-monsoon seasons.
• Projected a 1.0°C warming for Bangladesh by 2020 relative to 1971 levels.

Profound Impact of Artificial Neural Networks and Gaussian SVM Kernel on Distinctive Feature Set for Offline Signature Verification (May ‘12)

International Conference on Informatics, Electronics & Vision Mahabub Akram, Mahmud Ridwan, Tarif Ezaz, M Rashedur Rahman

System Overview

A comparative study of Artificial Neural Networks (ANN) and Support Vector Machines (SVM) for the purpose of verifying offline (static) signatures.

Feature Extraction

The system focuses on a distinctive feature set including:

• Global Features: Aspect ratio, center of gravity, and total area.
• Local Features: Gradient magnitudes and Histogram of Oriented Gradients (HOG).

Performance Results

• SVM with Gaussian Kernel: Achieved superior classification accuracy in identifying skilled forgeries.
• ANN Classifier: Demonstrated robust performance but required more intensive training data for convergence.

A Web-based Land Management System for Bangladesh (Dec ‘11)

14th International Conference on Computer and Information Technology DOI: 10.1109/ICCIT.2011.6164807

Project Goal

To design a framework that adds transparency and efficiency to the manual land management system in Bangladesh, reducing corruption and administrative delays.

Technical Framework

• Frontend: HTML5 Canvas for map viewing.
• Data Format: Scalable Vector Graphics (SVG) for digitizing paper-based land maps.
• Database: Searchable query interface for rapid land record retrieval (Khaatian and Daag numbers).

Implementation Details

The system bridges the gap between traditional paper records and digital accessibility, providing a user-friendly interface for both administrators and general citizens.