Face Recognition

# Simplified face recognition illustration
# Real systems use deep CNN models to compute face embeddings

import math

def fake_face_embedding(face_id):
    """
    In real systems, a CNN computes a 128-512 dimensional vector.
    We use tiny 3D vectors here to illustrate the concept.
    """
    embeddings = {
        "Alice":   [0.82, 0.11, 0.45],
        "Bob":     [0.21, 0.93, 0.12],
        "Charlie": [0.55, 0.48, 0.71],
        "Unknown": [0.90, 0.15, 0.40],  # close to Alice
    }
    return embeddings.get(face_id, [0, 0, 0])

def euclidean_distance(vec1, vec2):
    """Distance between two face embeddings."""
    return math.sqrt(sum((a - b)**2 for a, b in zip(vec1, vec2)))

# Known faces in database
database = {
    "Alice":   fake_face_embedding("Alice"),
    "Bob":     fake_face_embedding("Bob"),
    "Charlie": fake_face_embedding("Charlie"),
}

# Try to recognize an unknown face
def recognize_face(face_embedding, threshold=0.3):
    best_match = None
    best_distance = float("inf")
    
    for name, known_embedding in database.items():
        distance = euclidean_distance(face_embedding, known_embedding)
        if distance < best_distance:
            best_distance = distance
            best_match = name
    
    if best_distance <= threshold:
        return best_match, best_distance
    else:
        return "Unknown person", best_distance

# Test recognition
test_faces = {
    "Alice's face":     fake_face_embedding("Alice"),
    "Bob's face":       fake_face_embedding("Bob"),
    "Stranger's face":  fake_face_embedding("Unknown"),
}

print("Face Recognition Results:")
print()
for description, embedding in test_faces.items():
    match, distance = recognize_face(embedding)
    print(f"  {description}:")
    print(f"    Matched to: {match} (distance: {distance:.3f})")