Kinect Creative Studio

An industrial-grade, cross-platform interactive platform designed to breathe new life into Xbox Kinect V1 (Kinect for Windows, Model 1414/02c2) sensors or standard 2D Webcams. It merges generative ASCII text art pipelines with localized edge Machine Learning classification for real-time human gesture tracking.

📁 Repository Structure

kinect-creative-studio/
├── .gitignore
├── LICENSE                   # MIT Open-Source Legal License
├── README.md                 # Rich Setup and Deployment Guide
├── SPEC.md                   # Detailed Technical Architecture Specification
├── TROUBLESHOOTING.md        # Diagnostic and Failure Recovery Playbook
├── CHANGELOG.md              # Version Control & History
├── requirements.txt
├── config.json
├── install.sh
├── sanitize.sh               # Repository Hygiene Script
├── push.sh                   # Automated Quick Git Sync Tool
├── build.py
├── gestures_dataset.json
└── src/
    ├── __init__.py
    ├── main.py
    ├── gui_main.py
    ├── train_pipeline.py      # Independent Local AI Trainer
    ├── core/
    │   ├── __init__.py
    │   ├── driver_wrapper.py
    │   ├── gesture_engine.py
    │   └── gesture_model.pkl
    └── utils/
        ├── __init__.py
        ├── ascii_painter.py
        ├── trainer.py
        └── config_manager.py  # Non-volatile JSON Persistence Manager

🌟 Key Capabilities

• Hardware Agnostic Driver Matrix: Auto-detects Xbox Kinect depth channels on startup. Falls back seamlessly to generic RGB camera drivers (/dev/video0) using automated image luma downsampling if the sensor is disconnected.
• Interactive Neural Studio: An interface built with PySide6 allowing users to record positional patterns, train machine learning decision trees locally without GPU overhead, and map behaviors to system keys.
• Temporal Signal Deflickering: Uses rolling time-window queues to process matrices smoothly, preventing duplicate key entry fires caused by sensor tracking loss.

🛠️ Complete Environment Provisioning (Debian 12)

1. Automated Installation

Ensure your virtual environment is clean and execute the setup bundle wrapper script directly from the project root directory:

chmod +x install.sh
./install.sh

2. Active Desktop Initialization

Launch the unified graphical operations dashboard natively by invoking:

cd src
python gui_main.py

🧠 Step-by-Step Training Protocol

1. Configure System Mode: Switch the dropdown element labeled Analytical Engine Mode from GEOMETRIC to NEURAL.
2. Calibrate Static Floor (Repouso): Input the literal word Repouso into the text element. Stand clear in front of the lens in an upright resting pose. Press Start Recording and remain still for 5 seconds.
3. Capture Active Motion Action: Replace the string with your target gesture tag name (e.g., Wave or Swipe). Activate the trigger and move your right hand or body inside the spatial quadrant tracking zone.
4. Fit Model Tree: Press the blue Train Artificial Intelligence Model dashboard button. The state model updates in under 200ms.
5. Bind Trigger Targets: Select your target keyword string from the populated selector list, bind an OS execution shortcut key (e.g., space, right), and press Bind Keyboard Command.