Media Processor

A unified Python tool for video and image processing, combining video frame extraction, AI-powered image upscaling, and WebP to PNG conversion.

Features

• Video Frame Extraction: Extract frames from video files at specified intervals
• Image Upscaling: Upscale images using ONNX AI models (2x, 2.25x, 4x scaling)
• WebP to PNG Conversion: Convert WebP images to PNG format

Installation

1. Clone the repository:

git clone https://github.com/ebowwa/media-processor.git
cd media-processor

2. Install dependencies:

pip install -r requirements.txt

Usage

The tool provides a unified CLI interface for all operations:

Video Frame Extraction

Extract frames from a video file:

python cli.py video2frames input_video.mp4 -o output_frames/

Extract frames from all videos in a directory:

python cli.py video2frames videos_directory/ -o frames_output/

Options:

• -r, --rate: Frame extraction rate in seconds (default: 1 second)
• --keep-video: Keep original video files after extraction

Image Upscaling

Upscale a single image:

python cli.py upscale image.png -o upscaled_image.png

Upscale all images in a directory:

python cli.py upscale images_folder/ -o upscaled_folder/

Options:

• -m, --model: Choose upscaling model
  • modelx2: 2x upscaling (default)
  • modelx2_25_JXL: 2.25x upscaling
  • modelx4: 4x upscaling
  • minecraft_modelx4: 4x upscaling optimized for pixel art

WebP to PNG Conversion

Convert a single WebP image:

python cli.py webp2png image.webp -o image.png

Convert all WebP images in a directory:

python cli.py webp2png webp_folder/ -o png_folder/

Examples

1. Extract frames from a video every 2 seconds:

python cli.py video2frames movie.mp4 -r 2 -o movie_frames/

2. Upscale an image by 4x:

python cli.py upscale photo.jpg -m modelx4 -o photo_4x.jpg

3. Batch convert WebP images:

python cli.py webp2png downloads/ -o converted_pngs/

Supported Formats

• Video formats: MP4, AVI, MOV, MKV
• Image formats: PNG, JPG, JPEG, BMP, TIFF (for upscaling)
• Conversion: WebP to PNG

Requirements

• Python 3.7+
• OpenCV (cv2)
• NumPy
• Pillow (PIL)
• ONNX Runtime

Project Structure

media-processor/
├── cli.py                 # Main CLI interface
├── requirements.txt       # Python dependencies
├── models/               # AI upscaling models
│   ├── modelx2.ort
│   ├── modelx2_25_JXL.ort
│   ├── modelx4.ort
│   └── minecraft_modelx4.ort
├── src/
│   ├── video/
│   │   └── extractor.py   # Video frame extraction
│   ├── image/
│   │   └── upscaler.py    # Image upscaling
│   └── converters/
│       └── webp_converter.py  # WebP to PNG conversion
└── examples/             # Example files

License

This project consolidates functionality from:

• Video2Img
• upscale_img
• webp2png

Please refer to the original repositories for specific licensing information.