- The internet is a wonderful and informative platform. While we can leverage its potential to gain information about anything and everything, it also brings a lot of potential risk. The vulnarability while exploring the web brings a risk of showcasing us along with the children, with certain information which might bring a feeling of uncomfortability, destruction etc.
- Violent scenes are “scenes one would not let an 8 year old child see because they contain physical violence”.
This GitHub repository is intended to automatically classify the video content into violent or non-violent using the method of Deep Learning.
The visual and audio dataset information has been collected from Hollywood Movies and YouTube. The folder [annotations] contain the textual information about the violance annotations of the visual and audio modelarity. It has space-separated text format where each line in the former compiles to the structure: start-frame end-frame whereas, each line in the latter compiles to the structure: start-frame end-frame (concept-detail) Or start-second end-second (concept-detail)
The ipynb file- [DataAnalysis] contains the analysis of textual data or the informative dataset in the form of respective clear graph for the different datils of concept.
______________________________________________________________________________________________________________________
These links contain the final Jupyter Notebook on Google Collab.
Notebook link: https://drive.google.com/open?id=1w-hwquWq4Xl5MtDYPZhtrvctDLhCoyPX
Data: https://drive.google.com/open?id=1tSQrbXJ4dz-39xHSPYli3ePYZzKFl0T3 Note that, the dataset is under a license which prevents it from unauthorized sharing. It is provided here only for reference and must not be shared. Access will be removed after the hack is over.