arXiv subjects

Rafsan Jany

Publications and source records attributed to Rafsan Jany.

6 recordsLinked to original sources

arXiv2026

Twenty-Five Years of Air Quality in Bangladesh: Trends, Seasonality, and Spatial Pollution Regimes

Air pollution is one of Bangladesh's most pressing environmental problems, yet few studies have examined it at a national scale over a long time span. To describe trends, seasonality, geographic patterns, and pollution regimes, this study examines 25 years of hourly air quality data (2000-2025), covering over 3.19 million records, eight pollutants, and 103 cities monitored from 2022 onward following expansion from a single Dhaka station. The composite AQI series showed no significant long-term trend, an artifact of this network expansion rather than a real flattening of pollution, since Dhaka's own AQI rose steadily before the network grew. Seasonality was strong and highly significant, as AQI peaked in January (162.26) and fell to its lowest in July (61.47), tracking the dry and monsoon seasons. Geographically, southeastern coastal cities, including Teknaf and Bandarban, consistently recorded the lowest AQI levels. Dhaka remained the most polluted city, and air quality improved progressively with distance from the capital. K-means clustering of city-level AQI trajectories identified four distinct pollution regimes, ranging from cleaner coastal regions to a rapidly deteriorating Dhaka-Narsingdi cluster with an average increase of 1.84 AQI per year. Correlation analysis showed that PM2.5 and PM10 were the dominant contributors to AQI variation, and so combustion-related emissions are the primary target for air quality management and policy intervention.

physics.soc-ph↗

arXiv2026

Exploration-Driven Personalized Federated Reinforcement Learning via Intrinsic Motivation

Personalized Federated Reinforcement Learning (PFRL) takes a decentralized approach to storing and accessing information based on past experiences while keeping each client's data private during the learning of each client's policy. Many current methods for PFRL rely heavily on exploiting existing reinforcement learning reward signals to derive an optimal policy for each client, thereby neglecting exploration in non-stationary or sparse-reward environments. In this work, we introduce a new exploration-driven framework, Exploration-Driven Personalized Federated Reinforcement Learning via Intrinsic Motivation (EDPFRL-IM), that leverages an inherent curiosity-driven exploration at each client to promote local exploration and protect client privacy. Furthermore, to facilitate policy discovery via exploration in previously unexplored state spaces, clients add an intrinsic random network distillation (RND) signal to their extrinsic reward. Additionally, the server does not have access to clients' raw experiences or local gradient estimates; instead, the server sends global exploration priors and collects minimal novelty summaries from each client to enable both diverse and coordinated exploration among clients. Experiments in benchmark environments show that our framework outperforms average PFRL benchmarks in policy personalization and sample efficiency, primarily in delayed and sparse reward systems. Overall, EDPFRL-IM enables the integration of a flexible exploratory learning structure into federated reinforcement learning systems while preserving client privacy.

cs.LG↗

arXiv2026

UnDA: Unpaired Domain Alignment for Cross-Modal Knowledge Transfer in Medical Imaging

Multimodal based approaches often outperform single modality approaches in downstream tasks as the different modalities provide complementary information, yet acquiring paired clinical data remains a significant challenge in real world scenarios. While cross-modal knowledge distillation addresses this, existing methods often struggle with large modality gaps and the propagation of noise from uncertain source-domain predictions. To overcome these challenges, we propose UnDA, an anchor-guided framework for unpaired cross-modal distillation. Our approach introduces a backbone-agnostic Alignment Module that extracts semantically structured class tokens via an attention based pooling mechanism. To ensure robust knowledge transfer, we propose Uncertainty-Weighted Optimal Transport (UCT-OT), which dynamically weights feature-level alignment based on prediction confidence, effectively suppressing noisy supervision. Furthermore, a per-class ProtoNCE objective maintains stable prototype memories to enforce global discriminability across unpaired batches. Evaluations on representative segmentation tasks under strictly unpaired settings show consistent improvements in accuracy and boundary precision in the target modality, demonstrating that meaningful structural knowledge can be transferred across heterogeneous data sources without paired datasets.

cs.CV↗

arXiv2026

SleepExplain: Explainable Non-Rapid Eye Movement and Rapid Eye Movement Sleep Stage Classification from EEG Signal

Classification of sleep stages is one of the most important diagnostic approaches for a variety of sleep-related disorders. Electroencephalography (EEG) is regarded as a powerful tool for examining the association between neurological effects and sleep phases since it correctly identifies sleep-related neurological alterations. During Non-Rapid Eye Movement (NREM) and Rapid Eye Movement (REM) sleep phases, a number of nerve and bodily functions are affected and therefore hold an important role both in their functionalities. This work aims to classify NREM and REM sleep stages from sleep EEG data and present a noble SleepExplain model, an explainable NREM and REM sleep stage classification to explain its predictions. In this work, sleep stages were classified using Random Forest, XGBoost, and Gradient Boosting ensemble classification models. Overall, we obtained an accuracy of 92.54% (Random Forest), 94.25% (Gradient Boosting), and 94.30% (XGBoost). For explainable classification model, we utilized a game theoretic approach, SHAP (SHapley Addictive exPlanations) to offer a convincing explanation for the prediction.

cs.LG↗

arXiv2025

Balancing Accuracy and Efficiency: CNN Fusion Models for Diabetic Retinopathy Screening

Diabetic retinopathy (DR) remains a leading cause of preventable blindness, yet large-scale screening is constrained by limited specialist availability and variable image quality across devices and populations. This work investigates whether feature-level fusion of complementary convolutional neural network (CNN) backbones can deliver accurate and efficient binary DR screening on globally sourced fundus images. Using 11,156 images pooled from five public datasets (APTOS, EyePACS, IDRiD, Messidor, and ODIR), we frame DR detection as a binary classification task and compare three pretrained models (ResNet50, EfficientNet-B0, and DenseNet121) against pairwise and tri-fusion variants. Across five independent runs, fusion consistently outperforms single backbones. The EfficientNet-B0 + DenseNet121 (Eff+Den) fusion model achieves the best overall mean performance (accuracy: 82.89\%) with balanced class-wise F1-scores for normal (83.60\%) and diabetic (82.60\%) cases. While the tri-fusion is competitive, it incurs a substantially higher computational cost. Inference profiling highlights a practical trade-off: EfficientNet-B0 is the fastest (approximately 1.16 ms/image at batch size 1000), whereas the Eff+Den fusion offers a favorable accuracy--latency balance. These findings indicate that lightweight feature fusion can enhance generalization across heterogeneous datasets, supporting scalable binary DR screening workflows where both accuracy and throughput are critical.

cs.CV↗

arXiv2022

NFS: A Hand Gesture Recognition Based Game Using MediaPipe and PyGame

This paper represents a game which interacts with humans via hand gesture movement. Nowadays, apps like this seem rare, and there seems to be a window opening for this kind of application to be more prevalent and useful in the near future. This application is based on hand gesture movement instead of being dependent on a keyboard and mouse. The main issue was to figure out how to utilize machine learning to make this application work as it should be. First, two games were selected one with a traditional controller and another with hand gesture method. Then these two games based on the difficulty to use, fun elements, gameplay, and replayability were compared. Though the difficulty increases but the other three aspects improve significantly. After going through all of that a conclusion can be drawn that people are more likely to play a simple hand gesture-based game.

cs.HC↗