Institution profile

University of Florida

Academic institutionnorthamerica · us
Official website
Research library827linked papers
Opportunities0open roles
Selected work

Representative Papers

Eye-tracked Virtual Reality: A Comprehensive Survey on Methods and Privacy Challenges

May 23, 2023arXiv.org

This paper addresses privacy leakage risks arising from the correlation between eye-tracking data and visual stimuli in VR environments. We systematically survey full-stack VR eye-tracking technologies—from pupil detection and gaze estimation to cognitive modeling—published between 2012 and 2022, alongside their associated privacy threats. First, we establish the first cross-disciplinary survey framework bridging VR eye-tracking and privacy protection, identifying three privacy-centric research directions. Second, we propose a novel co-design paradigm integrating eye movement authentication with data anonymization, synergizing computer vision, human-computer interaction modeling, differential privacy, adversarial generation, and biometric encryption. Third, we clarify the technological evolution trajectory and privacy threat landscape, and introduce quantifiable evaluation metrics and an implementable defense roadmap. Our work provides both theoretical foundations and practical guidelines for developing secure and trustworthy VR systems. (149 words)

21 citationsRead paper

Ego-to-Exo: Interfacing Third Person Visuals from Egocentric Views in Real-time for Improved ROV Teleoperation

Jun 30, 2024arXiv.org

To address limited situational awareness in underwater ROV teleoperation caused by first-person (egocentric) vision, this paper proposes a geometry-driven, closed-form ego-to-exocentric view synthesis method that requires no training data and is cross-scene generalizable—enabling plug-and-play integration with existing monocular SLAM-based ROV systems. The approach synergistically combines real-time monocular SLAM pose estimation with lightweight 3D geometric modeling to reconstruct dynamic exocentric views under low-light conditions, supporting both 2-DOF indoor environments and 6-DOF underwater cave scenes. Subjective evaluations involving 15 operators demonstrate significant improvements in control accuracy and spatial situational understanding. Notably, it enables, for the first time, cave survey-line-guided navigation leveraging dynamically synthesized exocentric viewpoints. The core innovation lies in a zero-shot, geometry-prior-driven real-time view synthesis framework, overcoming the data- and scene-specific dependencies inherent in conventional learning-based methods.

9 citations1 influentialRead paper

Optimal Rate Region for Multi-server Secure Aggregation with User Collusion

Jan 11, 2026arXiv.org

This work addresses the problem of information-theoretically secure aggregation in a multi-server two-hop network, where up to $T$ users may collude with any subset of servers. Users communicate exclusively with their assigned servers, which then collaborate to recover the global sum. The study provides the first complete characterization of the optimal rate region for this setting, uncovering a fundamental trade-off between security and key efficiency. It demonstrates that a multi-server architecture substantially reduces the required randomness for secret keys. Within an information-theoretic security framework, leveraging linear key constructions and tight entropy bounds, the authors establish that the minimum user-to-server communication rate, inter-server communication rate, and individual key rate are each one symbol per input symbol, while the optimal source key rate is $\min\{U+V+T-2, UV-1\}$, where $U$ denotes the number of servers and $V$ the number of users per server.

2 citationsRead paper

MIRAGE: Multi-model Interface for Reviewing and Auditing Generative Text-to-Image AI

Mar 25, 2025

Harmful outputs from generative AI undermine its trustworthiness and societal deployment. To address this, we propose a public-facing, multi-model collaborative auditing framework implemented as a web platform that enables non-expert users to concurrently evaluate outputs from multiple text-to-image models, leveraging their personal identities and lived experiences to detect implicit biases and harmful content. Our contribution lies in the first design of a comparative multi-model auditing interface and a structured feedback mechanism, integrating heterogeneous model APIs, interactive visual comparison modules, and context-aware auditing forms. A preliminary user study (n=5) demonstrates that our approach significantly improves detection rates of harmful biases—particularly along gender, racial, and situational stereotyping dimensions. This work establishes a novel paradigm for democratized, interpretable, and participatory evaluation of generative AI systems.

2 citationsRead paper

A spatial interference approach to account for mobility in air pollution studies with multivariate continuous treatments

May 23, 2023

This study addresses exposure misclassification in causal inference of air pollution health effects due to individual mobility. Moving beyond the conventional measurement error paradigm, we propose a novel framework that formalizes population inter-regional movement as a geographic spillover interference process. We define and identify policy-relevant multivariate continuous-exposure causal quantities and rigorously prove that ignoring mobility induces systematic bias. Methodologically, we integrate mobile phone signaling, remote sensing, and ground-level monitoring data; estimate interference-adjusted causal effects via a Bayesian nonparametric model; and design a geographically weighted exposure reweighting strategy. Empirical analysis on the U.S. Medicare elderly population demonstrates that accounting for mobility substantially revises causal effect estimates of PM₂.₅ and NO₂ on mortality, improving exposure assessment accuracy and enhancing the interpretability and policy relevance of environmental health analyses.

2 citationsRead paper
Recent publications

Latest Papers