"If we knew what it was we were doing, it would not be called research, would it?"

Albert Einstein

Research Projects

Anomalous Activity Detection

This research develops a privacy-preserving, agent-based edge framework for anomaly detection in surveillance videos, integrating deep feature extraction and human-in-the-loop validation to enable efficient, real-time crime prevention. Key focus of this research:

🤖 AI-Driven Anomaly Detection 🔒 Privacy & Centralization Challenges 🧑‍⚖️ Human-in-the-Loop Decision Support

Publications: Patent 1 (granted - US11875566B1), Patent 2 (Submitted- Docket No. FIU.563)

Structural Damage Detection

Our key focus of this research:

🧠 AI-Enhanced Video Super-Resolution 🤖 Advanced Visual Language Model 📊 Validation & Performance

Publications: ICMLA '25

Federated Deep Learning for Edge Computing

We are developing privacy-preserving learning techniques in heterogeneous federated environments. Our key focus of this research:

📡 Edge Devices & IoT Data ⚙️ Resource Constraints FL 🔄 Federated Heterogeneity
Publications:

ICMLA '21, Springer Nature '22, ICCCN '22

Small Object Detection

Our key focus of this research:

🔍 Small Object Detection Challenge 📊 Benchmarking & Evaluation 🚀 Improved Detection Performance ⚡ Lightweight Models for Edge Devices

Publications:

IEEE Access (Submitted 2 papers)

Influence Operations

This research investigates how state-sponsored influence operations strategically deploy toxic language and differentiated emotional-rhetorical tactics across nations to manipulate discourse, amplify engagement, and advance geopolitical objectives. Key focus areas are:

😡😇 Emotional & Rhetorical Strategies 🧪 Toxic Language Analysis 🌐 Global Impact of IOs on Social Media 💻 Tools, Code & Transparency

Publications: ACM PETRA '25, ACM HT '25

Project name: Vision2FIM: Computer Vision Aided Curvilinear Topographic Feature Detection from LiDAR Data for Better Flood Inundation Modeling

Project name: Understanding Patterns and Prevention of Scamming by Impersonating Small Businesses

Project Title: Secure AI-Driven Crop Anomaly Detection: Integrating Federated Learning and Satellite Imaging for Precision Agriculture

Previous Projects

Agent-based Learning to Utilize Local Data for Anomalous Activity Recognition (Funded by the U.S. Department of Homeland Security)

We aim to detect anomalous events or suspicious activities such as assault, explosion, and shooting in surveillance videos.
We plan to improve the accuracy of decisions of human agents by reducing the manual work of monitoring of human agents.
We focus to provide better visualization to locate anomalous event and act accordingly.

House Detection from Aerial Images Using Faster RCNN

Demonstrate the effectiveness of Faster Region-based Convolutional Neural Network (Faster-RCNN) to detect buildings automatically from aerial images using Python programming language.

Key Notes:

Implementation of faster RCNN algorithm.
Data annotation using LabelMe.
Bounding-box information extraction from XML file.
Pre-trained ResNet50 for feature extraction.

Google Sites

Report abuse