Research Output

Publications

Peer-reviewed journal and conference work. Citation counts via Google Scholar.

Q1 Journal 2024 Cited by 22

DIA-VXNET: A framework for automated diabetic eye disease detection using transfer learning with feature fusion network

M. N. Hasan, M. E. R. Pial, S. Das, N. Siddique, H. Wang
Biomedical Signal Processing and Control, Vol. 100, Part C, 106907

A hybrid transfer-learning model fusing VGG16 and XceptionNet for diabetic eye disease detection. A transition block aligns mismatched feature shapes for effective concatenation, and the approach is benchmarked against 21 architecture combinations — achieving 99.76% classification accuracy.

Conference 2025

Evaluating Deep Learning Architectures for Actin Filament Segmentation Under Varying Noise Conditions in Simulated Cryo-ET Tomograms

M. E. R. Pial, F. N. Dehan, W. Wriggers, S. Sazzed
IEEE International Conference on Bioinformatics and Biomedicine (BIBM), pp. 7058–7065

A systematic study of how leading deep segmentation architectures hold up to increasing noise in simulated cryo-ET tomograms, with implications for robust actin filament extraction at low signal-to-noise ratios.

Conference 2025

SocialDemoExtract: A Tool for Extracting Self-Reported Age and Gender from Social Media Text

S. Sazzed, M. E. R. Pial, F. N. Dehan
IEEE International Conference on Bioinformatics and Biomedicine (BIBM), Wuhan, China, pp. 7645–7649

A natural-language tool that mines self-reported demographic signals — age and gender — from free-form social media text, supporting downstream population and health studies.

Conference 2025

Evaluating the Few-Shot Performance of Large Language Models for Classifying Clinical Risk Factors in Mental Health Text

S. Sazzed, R. S. Azade, F. N. Dehan, M. E. R. Pial
IEEE International Conference on Big Data (BigData), Macau, China, pp. 7714–7718

An empirical evaluation of few-shot prompting with large language models for identifying clinical risk factors in mental-health text, probing where in-context learning succeeds and fails on sensitive classification tasks.