Team picture
First eye track attempt
RF detector
Video StoryBoard
KnightHacks Leaderboard
Final UI Tracker
Looked Away

Inspiration

"Honesty is the Best Policy!" Or at least that's what we tell ourselves, but in reality, cheating has become a growing epidemic in the last few years with AI at the forefront of the problem. However, cheating on tests and exams has and will always be around so long as there are people wanting to take a shortcut. Although their shortcuts put every other honest person behind, and discourages real learning, so to bring back proper education we need to combat academic dishonesty.

What it does

Argos allows exam proctors to track students visual patterns and flag if they look away from their test too many times or for too long in a short amount of time. While also having the ability to spot iPhone, earbuds, and smartwatches both visible and not visible using both a camera and RF detector rig.

How we built it

We used an open-source AI face detector to help us track where students are facing and looking at all times. Then we then took those pre-determined positions and coded actions that check if this person is potentially cheating. As well as implementing an RF detector to check the radio-signals in the air that find out whether a device is being used in the area, and its approximate location.

Challenges we ran into

Changing AI models halfway into our build really through us off track for a bit, but it was necessary since the previous model lagged our app so bad we only reached around 3fps. Then we had trouble with our RF-detector because it kept picking up signals from across the building, along with previous signals that were already there.

Accomplishments that we're proud of

As a team we are so proud of all the hard work we were able to accomplish in 36 hrs, from the software in the code to the hardware we camped out to rent.

What we learned

"Perfection is overrated", sometimes you can't work out every bug, or implement every cool feature you want. Maybe that's because of time constraints, memory, or some other reason that stops you from perfecting your build, and maybe that ok too.

What's next for Argos

Taking Argos to the next step would lead to more features, such as overhead cameras, heat maps, ink sensors and more. So next time we would need better hardware, more budget, and more time to perfect our software and present it in front of real investors.