08 · Research · Computer vision
Steganographic QR communication
QR codes embedded inside ordinary-looking photographs, so links and messages can travel without exposing the encoded content.
I built a system that hides QR payloads inside natural images while keeping them readable after compression, noise, and everyday sharing conditions.
Hiding the fact that a message exists
A visible QR code announces that a message exists. The research question was whether a machine-readable signal could survive inside a photograph without making the image look engineered or suspicious to a human viewer.
Balancing image quality with detection
I trained and evaluated models that balanced two opposing goals: preserve image quality and keep the embedded code detectable. The pipeline tested resilience against common transformations like resizing, compression, and visual noise.
From experiment to mobile prototype
The prototype combined Python experimentation with CoreML and SwiftUI, turning the research into a mobile-facing system that could encode, carry, and recover hidden links or messages. Presented at KYSEF, the work drew research awards from four organizations.
U.S. Air Force
U.S. Marine Corps
U.S. Navy
General Electric