Touhid Alam

Dr. Touhid Alam

orcid

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
18 November 2025
Verified
Book Chapter
Open access
PDF file
TATouhid Alam
Touhid Alam
SAShusmita Anjum Aziz
Shusmita Anjum Aziz
SRSayedur Rahman
Sayedur Rahman
MMMohammad Sakib Mahmood
Mohammad Sakib Mahmood
Stroke imposes an immense burden on healthcare systems across the globe as a leading cause for mortality and permanent disablement. The condition occurs by an alteration in blood flow to the brain that results in neurological damage. As a result, early and accurate pre- diction of stroke risk is critical for timely intervention and prevention, which can significantly reduce its impact. Therefore, the primary objective of this study is to develop an ensemble-based Machine Learning (ML) model that can accurately predict the risk of stroke and identify the key factors contributing to it. To ensure optimal performance, data preprocessing was carefully implemented. Using a dataset sourced from Kaggle, comprising 5110 instances with 12 clinical and demographic features, the study employed multiple ensemble-based classifiers, including Decision Tree (DT), Random Forest (RF) classifier, Gradient Boosting (GB) classifier, AdaBoost classifier, Bagging classifier, and Extra Trees (ET) classifier. Among these models, the ET classifier achieved exceptional results, with 99.36% training accuracy and 99.43% testing accuracy. The minimal gap between training (99.36%) and testing (99.43%) accuracy indicates no significant overfitting, ensuring the model generalizes well to new data. Comprehensive evaluation metrics further highlighted the model’s performance, with precision (98.90%), recall (100%), F1-score (99.45%), Matthews Correlation Coefficient (MCC) (98.87%), AUC (1.0), RMSE (0.08%), and MSE (0.01%). Feature importance analysis conducted using the ET classifier identified age, average glucose level, and BMI as the most significant predictors of stroke. These results highlight the effectiveness of ensemble-based models in early stroke detection and risk stratification using demographic and clinical features.


21 September 2026 in Discover Artificial Intelligence
Verified
Journal Article
Open access
PDF file
TATouhid Alam
Touhid Alam
TLTze Hui Liew
Tze Hui Liew
MMMd. Kishor Morol
Md. Kishor Morol
DNDip Nandi
Dip Nandi
MJMd. Abdullah-Al Jubair
Md. Abdullah-Al Jubair
MRMashiour Rahman
Mashiour Rahman
Diabetic Retinopathy (DR) is one of the major causes of preventable blindness globally. Automated screening for DR is critical but is severely hindered by data scarcity and class imbalance. Real-world datasets, such as APTOS 2019, exhibit extreme class imbalance, where sight-threatening classes are statistically rare. Traditional augmentation fails to capture complex pathological features, while Generative Adversarial Networks (GANs) often suffer from mode collapse. Existing diffusion approaches typically operate in pixel space, limiting image resolution and requiring extreme computational resources. To address this, we introduce the Diabetic Retinopathy Latent Diffusion Synthesizer (DR-LDS), a class-conditional framework that synthesizes high-fidelity, \(512 \times 512\) fundus images natively deployable on consumer-grade hardware. By leveraging a fine-tuned Variational Autoencoder (VAE) for domain-adapted latent space compression, alongside an optimized U-Net, our method achieves superior anatomical realism with convergence in just 150 epochs (at 17 min 22 s per epoch). Extensive benchmarking demonstrates that DR-LDS outperforms state-of-the-art baselines, achieving a Fréchet Inception Distance (FID) of 8.05. When used to augment minority classes to a uniform target for the APTOS 2019 dataset across 11 deep learning architectures, our synthetic data significantly improved diagnostic accuracy by up to 15.25% (\(p < 0.001\)), enabling a lightweight SqueezeNet model to reach 95.60% validation accuracy and an F1 Score of 0.96. Moreover, rigorous data leakage analyses and Explainable AI (XAI) verify that DR-LDS learns genuine pathological biomarkers. This study establishes DR-LDS as a highly resource-efficient solution to the medical data bottleneck. The source code is available at: https://github.com/Touhid-Alam/DR-LDS

28 August 2026 in Applied Soft Computing
Verified
Journal Article
Closed access
TATouhid Alam
Touhid Alam
SHSultanul A. Hamim
Sultanul A. Hamim
JHJakir Hossen
Jakir Hossen
MMM. F. Mridha
M. F. Mridha