Committee Chair

Sartipi, Mina

Committee Member

Cox, Chris; Liang, Yu; Fisichella, Marco

Department

Dept. of Computer Science and Engineering

College

College of Engineering and Computer Science

Publisher

University of Tennessee at Chattanooga

Place of Publication

Chattanooga (Tenn.)

Abstract

Multi-target multi-camera tracking (MTMCT) is a cornerstone of intelligent transportation systems (ITS), enabling comprehensive monitoring and analysis of vehicle movements across distributed camera networks. Although significant progress has been made, fundamental challenges persist in data association, real-time processing, preservation of privacy, and natural-language user interaction. First, we propose a graph-based model that leverages similarity algorithms to enhance cross-camera object association. Our framework achieves state-of-the-art performance with an IDF1 score of 0.8166 on the CityFlow dataset for offline tracking while maintaining real-time capability at 14 FPS for online scenarios. Next, we present LaMMOn, an end-to-end framework integrating language models with graph neural networks. LaMMOn addresses data scarcity by generating synthetic embeddings, demonstrating competitive results in multiple datasets, including CityFlow (HOTA 76.46%) and TrackCUIP (HOTA 80.94%). To enable privacy-preserving tracking in large-scale deployments, we develop FLaMMOn, a federated learning framework incorporating federated elastic weight consolidation (FedCurv) and federated representation learning (FedRep). FLaMMOn outperforms centralized approaches with an IDF1 score of 76.04% while ensuring robust privacy guarantees. Finally, we introduce MACA, a large language multi-agent model that enables natural language for user queries for MTMCT (e.g., “Track black sedans moving from Camera 1 to Cameras 3 and 6 between 2 PM and 5 PM”). MACA achieves a HOTA score of 66.23% on our newly introduced Refer-CityFlow dataset. The comprehensive solutions presented in this dissertation enhance MTMCT systems through improved accuracy, scalability, privacy preservation, and user interaction capabilities. The proposed frameworks establish new benchmarks in performance while addressing critical real-world deployment challenges, paving the way for more effective and secure intelligent transportation systems.

Acknowledgments

I would like to express my deepest gratitude to my advisor, Professor Mina Sartipi, for her invaluable guidance, unwavering support, and dedication throughout my Ph.D. journey. Her invaluable advice and inspiration have profoundly shaped my research direction and academic growth. I cannot imagine going through this Ph.D. journey without her. I extend heartfelt thanks to my committee members, Professor Chris Cox, Professor Yu Liang, and Professor Marco Fisichella, for their expertise and insightful contributions. Their constructive feedback has been invaluable throughout the development of this dissertation. Special thanks to my collaborators and members of the research group for stimulating discussions and collaborative environment. Dr. Hoang Nguyen tirelessly shared his perspectives and experiences with me during my Ph.D. journey. Professor Ngan Le has offered unique perspectives that have greatly enriched my research. I also benefited from Dr. Nam Doan and other senior colleagues at CUIP for their academic mentorship. I would love to express my sincere appreciation for the active support from the Center for Urban Informatics and Progress (CUIP) and DOE, they generously helped bring many research ideas to life. I am also grateful to the Scruggs, the Semores, the Farmers, the Longs, Mr. Larry, and Mr. Dewey for keeping me in their prayers during my academic pursuit in the US. Many thanks are sent to Mr. Dat Pham, my dedicated life coach, for his sincere support and encouragement. To my family - Duoc Nguyen, Hue Le, Duc Nguyen, Giang Nguyen, Xuan Le, and Rau Ma. I owe immense gratitude to my father, who taught me to be generous and brave and how to be a man. He passed away last July, if only he could make it to my graduation party so I could see how proud he is. My mother has taught me to take risks and never give up despite any setbacks in life. My wife, Xuan Le, keeps my feet on the ground and my health in check. Her patience and loving care have been my foundation not only in this Ph.D journey but also in my life since we got married. Finally, I thank all my friends and colleagues in different research labs who made this journey memorable and enriching.

Degree

Ph. D.; A dissertation submitted to the faculty of the University of Tennessee at Chattanooga in partial fulfillment of the requirements of the degree of Doctor of Philosophy.

Date

5-2025

Subject

Computer vision; Federated learning (Machine learning); Graph algorithms; Intelligent transportation systems--Data processing; Privacy-preserving techniques (Computer science)

Keyword

Computer vision, graph neural networks, multi-target multi-camera tracking, large language models, federated learning, privacy preservation

Document Type

Doctoral dissertations

DCMI Type

Text

Extent

xxii, 165 leaves

Language

English

Rights

http://rightsstatements.org/vocab/InC/1.0/

License

http://creativecommons.org/licenses/by-nc-nd/3.0/

Date Available

5-31-2026

Download

Available for download on Sunday, May 31, 2026

Share

COinS