DeepFrame Studio

Welcome to our Django web app for video processing! This application provides a variety of features for manipulating and enhancing your video content. The main emphasis is on two powerful AI features: DF (DeepFake) detection using a fine-tuned MesoNet and video colorization via DeOldify.

Table of Contents

• Features
• Installation and Setup
• Usage
• Dependencies
• Contributing

Features

Video Processing

1. Video to Audio: Convert videos to audio effortlessly.
2. Change video format: Change videos format effortlessly.
3. Extract Frames from Video: Capture still frames from your videos.
4. Add Video to Audio: Merge video content with different audio tracks.
5. Reverse Video: Play your videos backward for a unique perspective.
6. Change Video Speed: Adjust video playback speed to suit your preferences.
7. Video Trimming: Trim your videos to the desired length.
8. Add Subtitles to Videos: Enhance your videos by adding subtitles.
9. Merge Multiple Videos: Combine multiple videos into a single file.

AI Features

10. DF Detection using MesoNet: Detect dense fog in videos using a fine-tuned MesoNet model.
11. Video Colorization via DeOldify: Apply advanced video colorization using the DeOldify model.

Installation and Setup

1. Clone the repository:

   git clone https://github.com/Humzafazal72/DeepFrame-Studio.git
   cd DeepFrame Studio
   

2. Download and install ffmpeg (You can follow this tutorial).

3. Import the Anaconda Virtual Environment using 'environment.yml' (You can follow this tutorial).

4. Install fastai:

   pip install fastai
      

5. Downloads the weights for DeOldify from this link:

6. Place the .pth weights in the folder 'models' along with mesonet's weights.

7. Run the project using the command:

   python manage.py runserver
   

Usage

Before any commercial use take permission from the owners of Deoldify and the creaters of CelebDFv2 dataset.

Dependencies

• Python 3.11.5
• Anaconda
• ffmpeg

Contributing

• Maheen Shahzad