TransConv-DDPM: Enhanced Diffusion Model for Generating Time-Series Data in Healthcare
Published in IEEE COMPSAC 2025, 2025
I presented our paper titled “TransConv-DDPM: Enhanced Diffusion Model for Generating Time-Series Data in Healthcare” at the IEEE 49th Annual Computers, Software, and Applications Conference (COMPSAC 2025), held in Toronto, Canada.
The presentation highlighted our proposed TransConv-DDPM framework — a transformer–convolution hybrid diffusion model designed to generate high-quality synthetic time-series data for healthcare analytics. The model enhances diversity, realism, and temporal consistency, supporting robust downstream tasks such as fall detection and activity recognition.
Citation:
Kabir, Md Shahriar, Sana Alamgeer, Minakshi Debnath, and Anne H. H. Ngu.
“TransConv-DDPM: Enhanced Diffusion Model for Generating Time-Series Data in Healthcare.”
In IEEE 49th Annual Computers, Software, and Applications Conference (COMPSAC), pp. 866–875, 2025.