AuraPass / Engineering case study
Offline identity and exam-access decisions, engineered as one accountable path.
AuraPass is a final-year software-hardware co-simulation prototype. It combines guided face enrollment, local eligibility data, seat allocation and a Proteus-simulated gate without depending on cloud services during an examination.
Controlled negative-face evaluation
The published numbers and what they actually prove.
The evaluation used 500 non-enrolled LFW source faces. Each was tested in 10 variants covering the original image, blur, brighter and darker exposure, high and low contrast, left and right rotation, and left and right shadow.
This is a false-grant stress test. It does not represent overall biometric accuracy, liveness performance or real campus deployment.
Download validation summary ↓- 5,000
- Total non-enrolled face attempts
- 3,793
- Denied after comparison
- 1,207
- Rejected by image-quality checks
- 0
- False grants in this controlled set
System architecture
A face match is necessary, but never sufficient.
AuraPass verifies identity, course eligibility, repeat entry, capacity and seat availability before it records the decision and sends the gate command.
- 01 Parse the course form and store guided face samples.
- 02 Verify identity, eligibility, repeat entry and capacity.
- 03 Reserve a seat and commit the decision.
- 04 Drive the LCD, LEDs, buzzer and simulated servo gate.
Hardware co-simulation
The decision reaches a visible, testable gate state.
A local bridge carries the committed access decision to the Proteus model, where the display, indicators, buzzer and servo response can be inspected in both grant and denial paths.
