Assignment_6.md

---

🔷 Shape & Contour Detection App

🎯 Objective

Build an interactive CLI or Streamlit tool that allows users to generate/draw basic geometric shapes, apply contour detection with polygon approximation, and view the annotated shapes with labels using OpenCV.

---

🔧 Instructions

1. Ask user:

"Which shape would you like to draw and detect?"

Choose one of the following options:

• 1. Triangle

• 2. Square

• 3. Rectangle

• 4. Pentagon

• 5. Hexagon

• 6. Circle

• 7. Ellipse

• 8. Star (5-point)

• 9. Regular Polygon (custom N sides)

---

🔺 If Option 1–9 Selected: Draw & Detect Shape

User Interface:

• Allow the user to choose:

  • Canvas size (e.g., 256–800 px).

  • Shape-specific parameters:

    • Circle → radius

    • Ellipse → axis_x, axis_y, rotation

    • Rectangle/Square → width, height, rotation

    • Regular Polygon → sides, radius, rotation

    • Star → outer radius, inner radius, rotation

Actions:

• Generate the binary image with the selected shape drawn in white.

• Preprocess with:

  _, thresh = cv2.threshold(img, 127, 255, cv2.THRESH_BINARY)

• Find contours:

  contours, _ = cv2.findContours(thresh, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)

• Approximate polygon curves:

  epsilon = 0.015 * cv2.arcLength(cnt, True)
  approx = cv2.approxPolyDP(cnt, epsilon, True)

• Classify shape based on:

  • Number of corners (len(approx)).

  • Aspect ratio (square vs rectangle).

  • Circularity & convexity (circle/ellipse vs irregular polygons).

• Annotate results with contours + labels:

  cv2.drawContours(annotated, [approx], 0, (0, 255, 0), 2)
  cv2.putText(annotated, shape_name, (x, y-10), cv2.FONT_HERSHEY_SIMPLEX, 0.6, (255, 0, 0), 2)

• Display both original drawn shape and detected shape with contours side by side.

• Offer Download option for results.

---

📏 Shape Classification Logic

• 3 corners → Triangle

• 4 corners → Square or Rectangle (check aspect ratio)

• 5 corners → Pentagon

• 6 corners → Hexagon

• >6 corners:

  • High circularity → Circle

  • Medium circularity → Ellipse

  • Concave polygon → “Concave Polygon (n)”

  • Else → “Polygon (n sides)”

📌 Note: circularity = (4π × area) / (perimeter²)

---

🖼️ If Option Selected: View Drawn vs Detected

• Left column: Original drawn grayscale shape.

• Right column: Annotated image with contours + detected label.

---

📄 Solution & 🌐 App

---