Diffusion Models for Controlled Architectural Image Generation
This project explores a novel hybrid diffusion model tailored for controlled architectural rendering. By integrating ControlNet with a hybrid ArchVisMix model, it ensures high-fidelity image generation of architectural elements. The framework optimizes architectural diffusion models by leveraging Flux model with ControlNet and LoRA training, enabling fine-tuned control over structural details.
More Detail
Unit Finder: ML-based Floor Plan Retrieval & Comparison
Unit Finder is a spatial feature-based unit plan retrieval system that identifies similar floor plans using boundaries, enclosures, and door locations. Integrated with architectural modeling software, it enables seamless data exchange through a user-friendly interface. The framework efficiently handles complex layouts with multiple enclosures and curves, ensuring adaptability across diverse architectural designs.
The Model Viewer application is a powerful visualization tool built on top of the Speckle platform, designed to enhance the management and interaction with architectural models. By leveraging Speckle’s robust framework, the app enables seamless integration with various CAD and BIM tools, providing a unified platform for visualizing, analyzing, and collaborating on complex architectural designs.
AI-Driven Innovations in Material Science
This research presents a smart AI-driven microscope system designed for the high-precision characterization of 2D materials, particularly transition-metal dichalcogenides (TMDs). This transformative system introduces a generative deep learning-based image-to-image translation method, enabling high-throughput and automated TMD characterization. By integrating deep learning algorithms, it can analyze and interpret data from multiple imaging and spectroscopic techniques, including optical microscopy, Raman spectroscopy, and photoluminescence spectroscopy, without the need for extensive manual analysis.
More Detail
Publications
Bhawsar, Shivani et al. AI-assisted Research for Transition Metal Dichalcogenides
Accepted Jenny Stanford Publishing (2025)
Bhawsar, Shivani. et al. Optimizing Diffusion Models for High-Fidelity Controlled Architectural Image Generation
Submitted (2025)
Bhawsar, Shivani. et al. Recent advances in ML and DL-enabled studies on transition metal dichalcogenides.
J. Phys. D: Appl. Phys. 58 073005 (2025)
Bhawsar, Shivani. et al. Deep learning-based multimodal analysis for transition-metal dichalcogenides.
MRS Bulletin 49, 1021–1031 (2024)
Bhawsar, Shivani. et al. Semantic Optimization for Fine-Tuned Image Tagging in Behance
Stevens Institute of Technology ProQuest Dissertations & Theses, 2023. 30695786