My research interests span signal processing, computer vision, and machine learning, with a focus on solving real-world problems through rigorous methods.
Research on channel state information reliability for vehicle-to-infrastructure communication in highway scenarios, leveraging passive intelligent surfaces for enhanced signal processing under mobility constraints.
Won both tracks of the international sclera segmentation competition, addressing biometric segmentation in the age of foundation models and large multi-modal models.
Track 1 - Foundation Models & VLMs: Integrated large multi-modal models into biometric segmentation pipelines using grounding approaches with textual and auxiliary prompts.
Track 2 - Label-Efficient Learning: Developed models using semi-supervised and self-supervised learning approaches, addressing data scarcity challenges in biometric imaging.