Inspiration
Who wants to be distributed by inappropriate content when watching a serie with your friends and family? Adult content should be accessed on deliberate search, not occurring in random situations.
What it does
This project aims to identify and blur inappropriate adult content to enable the user to pursue their activity. At the moment, the model detects inappropriate content and generates pop-up images to distract the user.
How we built it
We used a conventional neural network trained to detect content based on 5 categories: ['drawings', 'hentai', 'neutral', 'porn', 'sexy'].
Challenges we ran into
We use the Yolo algorithm to locate where this content is displayed, however, we did not manage to implement it. Blurring the specific part of the screen which is undesirable is also difficult, which is why we decided to pop images in a separate window instead.
Accomplishments that we're proud of
The detector works!
What we learned
Locating the content within the screen is hard and there is nothing existing to blur content in a window.
What's next for Adult content detector
Built With
- cnn
- python
Log in or sign up for Devpost to join the conversation.