- We have done many Youtube videos for school projects and we have found it difficult to blur out faces and license plates manually. We wanted to create this application to assistant and make that job a whole lot easier for not only future projects, but for anyone out there who needs a simple tool to assist them.
What it does
- This application will automatically blur out face profiles that it cannot identity, and any face profile that it recognizes (the people who are consenting to showing their face) will be left normal. The application also blurs out all license plates it recognizes.
How we built it
Challenges we ran into
- We as a team are new to image processing, getting it to work was a big enough of a challenge, even more so was optimizing the code so it didn't take tens of minutes to process ten seconds of footage. Another challenge we faced was scrambling to make sure it worked the best we could make it before the code was due, especially when things decide to not work properly near the end.
Accomplishments that we're proud of
- Getting deep learning to work on top of an initial layer of image processing is a huge accomplishment. Working together as a team to be able to achieve programming the base of this application is also something that we are very proud of.
What we learned
- Throughout the journey of this Hackathon, we learned a lot about image processing, computer vision, and web application integration that we did not previously have much experience in. We also all were able to look over each other's work and learn a bit more about languages that we were not as experienced in as others.
What's next for Aut0blur
- Optimizing, hardware acceleration and a dedicated desktop application in the near future. Hack K-State was only the beginning of this very simple yet useful application.