Project Details
[Return to Previous Page]Enhancing Airport Security Screening with Synthetic Data & AI
Company: Cignal LLC and Leidos, Inc. (co-sponsors)
Major(s):
Primary: CMPSC
Non-Disclosure Agreement: YES
Intellectual Property: YES
Cignal is partnered on this sponsorship with Leidos, a worldwide provider of aviation security equipment Cignal is the creator of the Cignal Engine, an advanced Generative AI (GenAI) image platform designed to enhance and improve the detection of dangerous and prohibited items at airport checkpoints. This technology offers a unique opportunity for university students to engage with cutting-edge AI and computer vision challenges within a real-world, high-stakes application. We propose a capstone project where a team of students would leverage Cignal Engine as a "wind tunnel" to explore and train a variety of novel algorithms aimed at enhancing current security screening capabilities. The project's primary goal is to empower students to develop and test innovative solutions for critical security tasks. This would include, but is not limited to, automatically segmenting luggage in complex X-ray CT scans to segment luggage items, identifying subtle or obvious anomalies that might indicate prohibited/dangerous items, building vision-language models to describe trends across images, and characterizing the contents of bags to distinguish between benign and concerning objects. Furthermore, a crucial element of the project is the exploration of novel visualization algorithms designed to assist security officers in efficiently examining X-ray CT images. This could involve developing techniques to highlight or render suspicious regions, fuse data from multiple detection models into an intuitive display, or present complex CT data in a way that significantly reduces officer fatigue and speeds up accurate decision-making. We envision this project culminating in the development of a suite of foundational security algorithms and practical visualization tools that demonstrate tangible improvements in security screening efficiency and accuracy. Students will gain invaluable experience working with a state-of-the-art synthetic data platform, allowing them to iterate rapidly and experiment with diverse machine learning and human-computer interaction approaches without the limitations or sensitive constraints of real-world data acquisition. This collaboration promises to provide students with a rich learning experience, preparing them for careers in AI, computer vision, and national security, while also contributing to the advancement of vital airport security technologies.
