http://dspace.daffodilvarsity.edu.bd:8080/handle/123456789/16 2026-04-05T15:03:28Z http://dspace.daffodilvarsity.edu.bd:8080/handle/123456789/16613 Assessing the Reliability of AI-Generated Medical Images: A Comparative Study with Real X-ray, CT, and MRI Scans Islam, Md Muhetul This paper presents a deep learning-based approach to differentiate between real medical images from AI-generated counterparts across X-ray, CT, and MRI images, addressing a critical challenge in healthcare diagnostics. Utilizing transfer learning with pre-trained models like InceptionV3, ResNet50, DenseNet121, VGG19, and MobileNetV2 on a 3,000 image dataset with 500 per class, the study achieves a peak accuracy of 0.82 with MobileNetV2, featuring an F1-score of 0.97 and near-perfect recall 1.00 for AI_MRI. InceptionV3 records 0.88, deeper models like ResNet50, DenseNet121, and VGG19 have accuracies below 0.60. It is relevant to engineering and medical standards to improve the reliability of the diagnosis. The framework follows IEEE 1012-2016 and DICOM standards. Limitations include the dataset's small size, leading to overfitting, like 0.9766 training vs. 0.8833 validation accuracy for InceptionV3. Future work proposes dataset expansion to 10,000 images, adversarial training, and edge device optimization to enhance diagnostic reliability and scalability in healthcare. Project Report 2025-09-16T00:00:00Z http://dspace.daffodilvarsity.edu.bd:8080/handle/123456789/16612 Detection of Rose Leaf Diseases Using Deep Learning Methods Lily, Anamika Hossain; Khan, Md. Walid Hassan The study revolves around automated detection and classification of rose plant diseases, a central aspect of high-end agricultural technology. Disease in rose plants is a serious threat to sustainable plant growth because manual detection is slow, imprecise, and often useless in preventing damage. The plant health loss not just downgrades its ornamental quality but also impacts the agric economy that relies on numerous individuals for their survival. Leaves being the plants' primary source of energy, any disease that infests them leaves the plant vulnerable. It is challenging to diagnose diseases in leaves because of their fragile appearance and environmental factors. To overcome this, deep learning techniques were employed, which were highly accurate at detecting the diseases from images. A dataset was made and augmented by data augmentation techniques such as rotation, flip, zoom, and brightness adjustment so that classes can be balanced and models can be generalized. The process involved included image preprocessing, data augmentation, and hyperparameter tuning with different CNN-based architectures. Preprocessing included resizing images, normalization, and format normalization. ResNet50, VGG19, Xception, and InceptionV3 were also tested for performance with four classes of rose leaf disease. Among them, the highest accuracy of 99.60% was recorded by InceptionV3 and it was well justified in its accurate classification. The approach is significantly promising for integration into automated systems to enable q Project Report 2025-09-17T00:00:00Z http://dspace.daffodilvarsity.edu.bd:8080/handle/123456789/16611 A Graph Neural Network Approach to Country-Level GDP Forecasting Saha, Parthib GDP forecasting is a major issue that requires accurate estimates on economic planning, policy formulation, trade policies and development objectives. Conventional statistical and machine learning models have delivered encouraging outcomes but have focused more on country level measures and have ignored the fact that countries are intricately linked through international trade. To overcome this shortcoming, this study introduces a GDP prediction method based on Graph Neural Network (GNN) which combines node-level economic signals and edge-level trade characteristics. The analysis makes use of the World Bank data (GDP, population, consumer price index, unemployment rate) and the BACI international trade database in order to create annual graphs, with the country being the node and the key trade relations the edges. Node features are based on economic indicators and edge features based on trade volumes of the ten most traded products worldwide. The Graph Attention Network v2 (GATv2) model incorporates multi-head attention, batch normalization and residual connections to learn country-level log-GDP representations and predictors. Through experimental analysis, the GATv2 model with edge features is shown to significantly outperform baseline machine learning models (linear regression, random forest) and previous graph-based models (GCN, GAT) across various evaluation metrics, including Mean Squared Error (MSE) and Mean Absolute Error (MAE), Root Mean Squared Error (RMSE) and R2 score. The findings underscore the need to integrate trade-based relational information in order to provide accurate macroeconomic forecasts. The paper will be relevant to the existing literature on graph-based economic modeling as it will illustrate the usefulness of attention-based GNNs in modeling dependencies in global trade. The given framework can be applied to other macroeconomic indicators and provide substantial information to policymakers, economists, and researchers in order to create data-driven economic policies. Project Report 2025-09-16T00:00:00Z http://dspace.daffodilvarsity.edu.bd:8080/handle/123456789/16610 Bengali social media comments classification and toxicity detection using advanced machine learning algorithms Rafin, Rayhan; Shibly, Mohammad Sohaib Islam This project presents a system for classifying topics and sentiments from Bengali comments using both machine learning and deep learning algorithms. Detecting toxic behavior and hate speech is a primary usage of this system. All steps taken to complete this experiment were done maintaining widely accepted standards. One of the challenges of this experiment was working with a low-resource language like Bengali that required custom and extensive preprocessing and model design. All the trained models were evaluated using standard metrics. Aside from the technical aspects, this project also focuses on improving online safety and moderation for Bengali language. A comment analysis tool for web view is also developed as part of this project. Project Report 2025-09-16T00:00:00Z