From developing an Educational Security Operations Center for hands-on cybersecurity training to applying machine learning for smart parking solutions and designing autonomous IoT systems like an aeroponic gardening tower, Information Technology students apply their skills to solve real-world challenges.
In their final year, students complete a two-course capstone sequence that mirrors the full lifecycle of a technology project. In IT 444: Capstone Project Design, teams identify a real-world problem and develop a proposed IT-based solution. In IT 445, those designs are implemented and integrated as students manage the project, collaborate as a team and deliver a functional solution.
Projects span areas such as cybersecurity, networking, web and mobile development and systems integration. Students define requirements, build and test solutions and communicate their results through professional reports presentations and prototypes.
The CISE Showcase, held in April, highlights this work as students present their projects and share the impact of their solutions with faculty, peers and industry partners.
Below are this year’s Information Technology capstone projects.
STUDENTS: Andrew Tran and Landon Tran
ADVISOR: Dr. Ahmad Salman
SkyVision is about developing a drone system that can help people in emergency situations by detecting obstacles and enabling safer aerial inspections.
The team is building and testing the system so it can eventually be used in real-world field conditions. Current work includes improving how the drone reacts to objects, checking and refining its detection accuracy, troubleshooting flight controls, and preparing for controlled flight tests.
The goal is to create a drone system that can help with tasks such as surveying damaged areas, identifying hazards, and providing responders with a better view of a scene from above.
STUDENTS: Michael Simmons, Jacob Massa and Vanmartin Leang
ADVISOR: Dr. Suk Jin Lee
As generative AI and deepfake technology have become more commonplace in everyday life, they are increasingly integrated across sectors such as education, accessibility, and entertainment. While these technologies offer significant benefits, they also create opportunities for malicious use, including deception, fraud, and the spread of disinformation.
The objective of this project is to develop software capable of detecting artificial intelligence with minimal false positives or negatives, while remaining efficient enough to run on an IoT edge device, such as Nvidia Orin NX. This proposed system is designed for use during video calls, such as Zoom.
This method is based on a simple video-calling application built with WebRTC and a Flask web server written in Python. Incoming video from the remote stream is processed frame by frame using OpenCV and analyzed by a classification model integrated with Vision Transformers. The model was trained using a variety of publicly available datasets through Google Colab.
The system's output determines whether the incoming video is authentic. Our final data evaluates the model's effectiveness in detecting AI-generated content and demonstrates its potential to help reduce fraud and limit the spread of AI-driven misinformation.
STUDENTS: Connor Prince, Nathan Park and Dylan Dela Rosa
ADVISOR: Dr. Suk Jin Lee
Facial recognition access control systems are becoming increasingly common in smart homes, offering convenience, reliability and hands-free operation. However, many existing systems remain vulnerable to spoofing attacks, such as printed photos, replayed videos or deepfake media. These vulnerabilities can compromise reliability in real-world scenarios.
This project addresses these challenges while preserving the convenience of facial recognition for residential access control. It focuses on the design, implementation and evaluation of a secure facial recognition system.
The system combines modern deep learning techniques with traditional computer vision algorithms and an active liveness verification approach. It evaluates how a deep learning-based facial embedding model, such as FaceNet, performs compared with a traditional Local Binary Patterns Histogram (LBPH) method.
A key contribution of this project is the use of an interactive challenge-response liveness-detection process, which verifies the physical presence of a user to prevent spoofing.
STUDENTS:
ADVISOR: Livia Griffith
Short paragraph about research
STUDENTS:
ADVISORS: Dr. Prajakta Belsare and Samy El-Tawab
Short paragraph about research
STUDENTS:
ADVISOR: Dr. Prajakta Belsare
Short paragraph about research
STUDENTS:
ADVISORS: Livia Griffith, Dr. Bryan Cage and Dr. Suk Jin Lee
Short paragraph about research
