Rutuja Patil
Computer systems engineering
Hometown: Pune, Maharashtra, India
Graduation date: Fall 2025
Additional details: Transfer student
FURI | Spring 2025
Optimizing Video Question Answering for Traffic Monitoring Systems
In contrast to large language models (LLMs) excelling in textual tasks, significant challenges remain in processing visual data. Video Question Answering (VQA) models exhibit low performance due to limited datasets, a lack of spatiotemporal understanding (time-space relationships), and insufficient contextual depth. This research aims to enhance VQA models for traffic monitoring systems by addressing these challenges. The study evaluates existing models on their performance with video inputs, optimizes VQA architectures with attention-based mechanisms tailored to traffic scenarios, and generates annotated datasets specific to traffic monitoring. By leveraging a comprehensive dataset from Argos Vision Inc., the research focuses on detecting traffic participants, analyzing events such as congestion and accidents, and adapting to real-world intersection dynamics. Overall, this research seeks to improve the standing of VQA models in the traffic monitoring domain, promoting a deeper understanding of complex spatiotemporal dynamics and enabling accurate, contextually relevant answers for real-time traffic analysis.
Mentor: Bharatesh Chakravarthi
It’s hip to be square.
Students presenting projects at the Fulton Forge Student Research Expo are encouraged to download this personal QR code and include it within your poster. This allows expo attendees to explore more about your project and about you in the future.
Right click the image to save it to your computer.