Skip to content

Repository navigation

README.md

Latest commit

 

History

History
145 lines (111 loc) · 5.51 KB

README.md

File metadata and controls

145 lines (111 loc) · 5.51 KB
Raw

Real-Time Shape & Finger Detector

A Python application that uses your computer's webcam to detect and classify geometric shapes and count fingers in real-time using OpenCV.

🚀 Features

Shape Detection

  • Real-time Detection: Processes live video feed from your webcam.
  • Shape Classification: Identifies Triangles, Squares, Rectangles, Pentagons, Hexagons, and Circles.
  • Visual Feedback: Draws contours around detected shapes and labels them with their name.
  • Noise Reduction: Filters out small contours to prevent flickering and false detections.
  • Traditional CV Approach: Uses contour approximation and geometric properties instead of heavy machine learning models.

Finger Detection

  • Hand Detection: Uses HSV color space skin detection to identify hand regions.
  • Finger Counting: Estimates finger count using convex hull and convexity defects analysis.
  • Visual Feedback: Draws hand contour, convex hull, and finger gap points.
  • Real-time Overlay: Displays finger count on the video feed.

General Features

  • FPS Counter: Real-time frame rate display.
  • Toggle Controls: Enable/disable shape detection (S) and finger detection (F) independently.
  • Information Panel: On-screen display showing FPS, finger count, and detection modes.

🛠️ Technologies Used

  • Python 3.x - Programming language
  • OpenCV (opencv-python) - Computer vision library for image processing
  • NumPy - Numerical computing for array operations
  • Math & Time - Standard libraries for calculations and timing

📦 Installation

  1. Clone or create the project folder:

    mkdir shape_detector
cd shape_detector

  2. Create a virtual environment:

    python -m venv venv

  3. Activate the virtual environment:

    • Windows: 
      .\venv\Scripts\activate
    • macOS/Linux: 
      source venv/bin/activate

  4. Install dependencies:

    pip install -r requirements.txt

💻 How to Run in VS Code

  1. Open the shape_detector folder in VS Code.
  2. Open a new terminal (Ctrl + Shift + `` ``).
  3. Ensure your virtual environment is activated (see installation steps).
  4. Run the program: 
    python main.py

📷 Webcam Permissions

  • Windows: Ensure that "Camera access" is turned on in Settings > Privacy & security > Camera.
  • macOS: When you run the script, macOS may prompt you to allow VS Code or Terminal to access the camera. Click OK.
  • Linux: Ensure your user is part of the video group: sudo usermod -aG video $USER.

🔍 How Shape Detection Works

The program follows these image processing steps:

  1. Grayscale Conversion: Converts the colorful camera feed to grayscale to simplify processing.
  2. Gaussian Blur: Smooths the image to reduce high-frequency noise.
  3. Canny Edge Detection: Identifies the edges of objects in the frame.
  4. Contour Finding: Connects the edges to find closed shapes.
  5. Contour Approximation (cv2.approxPolyDP): Simplifies a complex contour into a polygon with a smaller number of vertices. This is the key to identifying shapes:
    • 3 vertices → Triangle
    • 4 vertices → Square or Rectangle
    • 5 vertices → Pentagon
    • 6 vertices → Hexagon
    • Many vertices → Circle
  6. Geometric Analysis: For 4-sided shapes, the program calculates the aspect ratio (width/height) to distinguish between a square (ratio ≈ 1) and a rectangle.

🖐️ How Finger Detection Works

  1. Color Space Conversion: Converts BGR frame to HSV for better skin color segmentation.
  2. Skin Mask Creation: Applies HSV threshold range to isolate skin-colored pixels.
  3. Morphological Operations: Opening and closing to remove noise and fill gaps in the mask.
  4. Hand Contour Detection: Finds the largest contour in the mask (assumed to be the hand).
  5. Convex Hull: Computes the convex hull of the hand contour.
  6. Convexity Defects: Finds indentations in the hull (gaps between fingers).
  7. Angle & Depth Analysis: Filters defects by angle (< 90°) and depth (> 10px) to count finger gaps.
  8. Finger Count: Finger count = finger gaps + 1 (clamped to 0-5).

⌨️ Keyboard Controls

Key Action
Q Quit the application
S Toggle Shape Detection (ON/OFF)
F Toggle Finger Detection (ON/OFF)
R Reset FPS timer

🛠️ Troubleshooting

  • Webcam not opening:

    • Ensure no other app (Zoom, Teams, etc.) is using the camera.
    • Try changing cv2.VideoCapture(0) to cv2.VideoCapture(1) in main.py if you have multiple cameras.

  • Shapes not detected:

    • Use a simple, contrasting background (e.g., a white piece of paper on a dark table).
    • Ensure there is good lighting.

  • Too much noise:

    • Adjust the MIN_SHAPE_AREA constant in main.py to ignore smaller objects.

  • Finger detection not working:

    • Ensure your hand is clearly visible with good lighting.
    • Skin detection works best with standard lighting conditions.
    • Adjust SKIN_LOWER and SKIN_UPPER HSV ranges in main.py for different skin tones.

📁 Project Structure

shape_detector/
├── main.py          # Main application with shape & finger detection
├── requirements.txt # Python dependencies
└── README.md        # This file

🤝 Contributing

Feel free to fork this repository and submit pull requests for improvements.

📄 License

This project is open source and available under the MIT License.