Image Analysis and Metadata Generator

A Python-based tool for analyzing images and generating structured metadata in CSV format. This tool processes images to extract meaningful information and organizes it into a structured format for easy integration with other systems.

Features

• Image Analysis: Processes images to extract visual features and characteristics
• Metadata Generation: Creates structured data including:
  • Image titles
  • Alt text descriptions
  • SEO-optimized descriptions
  • Scene classification (indoor, outdoor, abstract, etc.)
  • Visual cohesion scores
  • Relevant tags and categories
• CSV Export: Saves all generated metadata in a structured CSV format
• Batch Processing: Handles multiple images in a single run
• Progress Tracking: Shows real-time progress of image processing

Requirements

• Python 3.8+
• PyTorch 2.0+
• CUDA-capable GPU (recommended)
• 8GB+ RAM

Installation

1. Clone the repository:

git clone https://github.com/tatianathevisionary/poster-analysis.git
cd poster-analysis

2. Create and activate a virtual environment:

python -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate

3. Install dependencies:

pip install -r requirements.txt

4. Test the installation:

python test_installation.py

Usage

1. Place your images in the input/ directory
2. Run the analysis script:

python inference_ram.py

3. Check the results:

• Generated metadata will be saved in output/metadata.csv
• Logs are available in poster_analysis.log

CSV Output Format

The generated CSV file (output/metadata.csv) includes the following columns:

• image_path: Path to the processed image
• title: Generated title for the image
• alt_text: Descriptive alt text for accessibility
• seo_description: SEO-optimized description
• tags: Comma-separated list of relevant tags
• scene_type: Classification of the scene (indoor, outdoor, abstract, etc.)
• cohesion_score: Numerical score indicating visual harmony (0-1)

Directory Structure

poster-analysis/
├── input/              # Place your images here
├── output/             # Generated metadata and results
├── ram/               # RAM model files
├── pretrained/        # Pretrained models
├── inference_ram.py   # Main analysis script
├── test_installation.py
├── requirements.txt
├── README.md
└── LICENSE

Contributing

1. Fork the repository
2. Create a feature branch
3. Commit your changes
4. Push to the branch
5. Create a Pull Request

License

This project is licensed under the MIT License - see the LICENSE file for details.

Support

For support, please open an issue in the GitHub repository or contact support@metaposters.ai