FakeDetect

Inspiration

With the growing need for awareness in our age of digital media, data manipulation is ever so prevalent. FakeDetector arrises from that unexploited ability to manipulate images for negative purposes.

What it does

FakeDetect at its heart is machine learning. We used a Jupyter notebook to take images of faces and check if the images have been manipulated.

How we built it

We built FakeDetect with python and Tensorflow. The front-end is a simple website layout to comfort users.

Challenges we ran into

Nobody on our team knew how to use Tensorflow so we spent much of the weekend learning. Additionally, there were many distractions at Treehacks such as all the cool workshops and events that took up a ton of our time.

Accomplishments that we're proud of

I am happy to say we have a basic product that is described on a website and has one implementation.

What we learned

Tensorflow and machine learning gets a lot more complex when working with large data sets such as images. Also, we do not have any training in machine learning or data analysis.

What's next for FakeDetect

FakeDetect was a great starter project to learn machine learning and probably will not be directly continued. Instead, we will take our newfound knowledge to explore other challenges with more direction.