Inspiration
As CS students, we have many times encountered ourselves clicking away from the algorithms videos we intended to watch. When willpower isn't enough, NSFW can come in to help out.
What it does
Using both deep learning and categorical analysis of Youtube videos, our extension selectively blocks irrelevant materials with up to a 98% accuracy. The extension allows the user to select what filters they wish to apply, such as "Click-bait", "Computer Science", "Economics", etc.
How we built it
The deep learning algorithm was built using keras on top of tensor flow, convenient Python libraries for constructing recurrent neural networks. The front end was built using the Chrome extension API. This utilizes mostly javascript with HTML and CSS being used occasionally. Communication between machine learning algorithm and front end was attempted with a REST API, which ended in a semi-functional state (unable to pass all information required).
Challenges we ran into
Sleep.
Accomplishments that we're proud of
What we learned
What's next for NSFW
Built With
- chrome
- flask
- javascript
- keras
- python
- tensor-flow
Log in or sign up for Devpost to join the conversation.