iHARP Research Assistant
Email: mhassan4@umbc.edu
See me on LinkedIn
Research Interests
- Artificial Intelligence
- Machine Learning
- Computer Vision
Short Biography
Muhammad Behroze Hassan is a graduate student in Information Systems at the University of Maryland, Baltimore County. He holds a Bachelor’s degree in Computer Science, specializing in Machine Learning, Deep Learning, and Computer Vision. Right now, under the supervision of Dr. Vandana Janeja, his research focuses on applying AI-driven techniques for enhanced glacial layer detection and annotation validation in ice-penetrating radar data. Behroze’s previous projects include developing AI-powered smart glasses for visually impaired individuals and implementing real-time face mask detection using machine learning. His goal is to leverage advanced AI techniques to address complex environmental challenges.
Research Summary
Behroze is currently working on developing automated annotation techniques for ice-penetrating radar (IPR) imagery to extract internal ice layers critical for ice-sheet modeling. A key focus of his research is designing a ground-truth-independent validation metric to evaluate the consistency and accuracy of these annotations, both manual and automated, without relying on expert-labeled data. By leveraging radargram intersections and computing directional similarity of layers, this approach helps identify annotation errors (including hallucinated layers) and supports more robust, scalable layer detection. This work contributes toward improving the quality of cryospheric data products used in climate modeling, ice flow simulation, and AI-based training pipelines.
Last updated 11 June 2025