Scene Sync - Beta 1

A Python application that matches film photos with scene photos using ORB (Oriented FAST and Rotated BRIEF) feature detection and matching.

Beta 1: to verify the feasibility of scene matching via ORB

Features

• Command Line Interface: Process photos directly from the terminal
• Web Interface: User-friendly browser-based interface
• ORB Feature Detection: Advanced image matching using OpenCV
• Flexible Input: Support for various image formats

Project Structure

scene-sync/
├── app/
│   ├── __init__.py
│   ├── core/
│   │   ├── __init__.py
│   │   ├── matcher.py          # ORB matching logic
│   │   ├── utils.py            # Utility functions
│   │   └── verifier.py         # For test verification
│   ├── cli/
│   │   ├── __init__.py
│   │   └── commands.py         # CLI commands
│   └── web/
│       ├── __init__.py
│       ├── routes.py           # Flask routes
│       ├── templates/          # HTML templates
│       └── static/             # CSS, JS, images
├── film-photos/                # Input folder for film photos
├── scene-info/                 # Input folder for scene photos
├── output/                     # Output folder for results
├── truth/                      # Truth folder for true pair datasheet storage
├── requirements.txt
├── main.py                     # Main application entry point
└── README.md

Setup

1. Install Python Dependencies:

   pip install -r requirements.txt

2. Create Required Directories:

   mkdir -p film-photos scene-info output

3. Add Your Photos:

   • Place film photos in film-photos/ (organized in subfolders)
   • Place scene photos in scene-info/ (organized in subfolders)

Usage

Command Line Interface

# Match photos from specific folders
python main.py match --film-folder film-photos/folder1 --scene-folder scene-info/folder2

# List available folders
python main.py list-folders

# Run with custom output directory
python main.py match --film-folder film-photos/folder1 --scene-folder scene-info/folder2 --output output/results.csv

Web Interface

# Start the web server
python main.py web

# Then open http://localhost:5001 in your browser

Input Requirements

• Image Formats: JPEG, PNG, BMP, TIFF
• Folder Structure: Each folder should contain only image files
• File Names: Must be unique within each folder
• Image Quality: Higher resolution images work better for matching

Output

The app generates a CSV file with columns:

• film_photo: Name of the film photo file
• scene_photo: Name of the matched scene photo file
• confidence_score: Matching confidence (0-1)

Configuration

You can adjust matching parameters in app/core/matcher.py:

• MAX_FEATURES: Maximum number of features to detect
• GOOD_MATCH_PERCENT: Threshold for good matches
• MIN_MATCHES: Minimum matches required for a valid match