The rise of AI and deep learning technology has lead to new technological advancements in society. But it also poses a critical threat to security and authenticity. It holds the power to impersonate important figures with absolute perfection. Imagine a scenario where an AI generated president authorized an attack that could lead to world war three? However, we can fight deep fakes using the same technological advancements.
What it does
Quick Live is the new age of video authentication. Using the same technology that brought you Bitcoin, Quick Live stores important metadata at the time of recording. Then, it stores a transaction of that recording in a dedicated block-chain. That's right, decentralized records means a seal of authenticity for your important videos.
All you need to get started is to open up the the browser to our secure dashboard. Using a webcam, record your true self as our powerful back-end hashes your video frames for the block-chain. Will let you know when your video is fully processed.
How we built it
Challenges we ran into
Believe it or not, the challenge was to get the webcam frames delivered to the backend. Because of the latency factor that comes from transferring images twenty four times per second, it was hard to optimize a real-time method for up-streaming.
Another challenge was processing the frames received and using the block-chain API. When we switched over to Flask, it was rather difficult to run multi-threaded processes that are required to hash frames.
Accomplishments that we're proud of
For the first time in a hackathon, we used a web based front-end using Bootstrap. As a result, our project looks professional and polished. It also makes it accessible for most users, especially on mobile.
The video streaming itself was a grand accomplishment that was challenging to implement. It pushed us to use other web technologies that we were unfamiliar with.
What we learned
This hackathon was very education for all of us. In addition to learning more about proper git collaboration, we dived into a new back-end framework (Flask) that was unfamiliar. It was similar to Express.js as it used routes and rendering templates.
On the other hand, we learned about web-sockets for the first time and how to provide consistent communication between the server and client. The video data needed to be optimized for sending at high speeds.
Finally, we learned about the fascinating word of block-chain transactions and how to write our own using a simple API. All being done in the comfort of a Python interpreter,
What's next for Quite Live
As we only had the time to create the general process for processing videos, we would add more CRUD functionality and user authentication to keep track of recent video transactions.