import cv2 import numpy as np import matplotlib.pyplot as plt from matplotlib.patches import Patch, Circle, Ellipse, Polygon import os import urllib.request import tempfile import shutil # Get the directory where this script is located script_dir = os.path.dirname(os.path.abspath(__file__)) # Download the face landmarker model if not present model_path_src = os.path.join(script_dir, "face_landmarker.task") if not os.path.exists(model_path_src): print("Downloading MediaPipe Face Landmarker model...") model_url = "https://storage.googleapis.com/mediapipe-models/face_landmarker/face_landmarker/float16/1/face_landmarker.task" urllib.request.urlretrieve(model_url, model_path_src) print("Model downloaded successfully!") # Copy to temp dir (MediaPipe cannot handle Unicode in paths) model_path = shutil.copy2(model_path_src, tempfile.mkdtemp()) # Import MediaPipe import mediapipe as mp from mediapipe.tasks import python from mediapipe.tasks.python import vision def detect_face_landmarks(image_rgb, model_path): """Detect facial landmarks using MediaPipe.""" mp_image = mp.Image(image_format=mp.ImageFormat.SRGB, data=image_rgb) base_options = python.BaseOptions(model_asset_path=model_path) options = vision.FaceLandmarkerOptions( base_options=base_options, num_faces=1 ) detector = vision.FaceLandmarker.create_from_options(options) result = detector.detect(mp_image) detector.close() if not result.face_landmarks: return None height, width = image_rgb.shape[:2] landmarks = [] for lm in result.face_landmarks[0]: x = lm.x * width y = lm.y * height landmarks.append((x, y)) return landmarks def get_landmark_region(idx): """ Return the facial region name and color for a landmark index. MediaPipe Face Mesh has 478 landmarks organized into regions. The exact indices are based on the MediaPipe face mesh topology. """ # Define regions with colors (approximate index ranges) if idx >= 468: # Irises return "Irises", "#00FFFF" # Cyan elif idx in [10, 338, 297, 332, 284, 251, 389, 356, 454, 323, 361, 288, 397, 365, 379, 378, 400, 377, 152, 148, 176, 149, 150, 136, 172, 58, 132, 93, 234, 127, 162, 21, 54, 103, 67, 109]: return "Face Oval", "#00FF00" # Green elif idx in range(33, 134) or idx in range(263, 364): return "Eyes", "#FF8000" # Orange elif idx in range(17, 33) or idx in range(133, 168) or idx in range(168, 198): return "Eyebrows", "#0000FF" # Blue elif idx in range(1, 17) or idx in range(195, 263): return "Nose", "#FFFF00" # Yellow elif idx in [61, 146, 91, 181, 84, 17, 314, 405, 321, 375, 291, 409, 270, 269, 267, 0, 37, 39, 40, 185, 78, 191, 80, 81, 82, 13, 312, 311, 310, 415, 308, 324, 318, 402, 317, 14, 87, 178, 88, 95]: return "Lips", "#FF0000" # Red else: return "Face Surface", "#808080" # Gray # Main execution if __name__ == '__main__': # Load the sample face image image_path = os.path.join(script_dir, "sample_face.jpg") image_bgr = cv2.imread(image_path) if image_bgr is None: print(f"Error: Could not load image from {image_path}") exit(1) # Convert BGR to RGB image_rgb = cv2.cvtColor(image_bgr, cv2.COLOR_BGR2RGB) height, width = image_rgb.shape[:2] print("Detecting facial landmarks...") landmarks = detect_face_landmarks(image_rgb, model_path) if landmarks is None: print("No face detected!") exit(1) print(f"Detected {len(landmarks)} landmarks") # Create visualization fig, axes = plt.subplots(1, 2, figsize=(16, 8)) # Left: Original image with landmarks axes[0].imshow(image_rgb) axes[0].set_title('Facial Landmarks (478 points)', fontsize=14, fontweight='bold') axes[0].axis('off') # Draw landmarks with color-coded regions region_colors = {} for idx, (x, y) in enumerate(landmarks): region, color = get_landmark_region(idx) region_colors[region] = color # Vary marker size by region importance if region == "Irises": size = 40 elif region in ["Eyes", "Lips"]: size = 15 else: size = 8 axes[0].scatter(x, y, c=color, s=size, alpha=0.8, edgecolors='white', linewidths=0.3) # Add legend legend_patches = [Patch(facecolor=color, label=region) for region, color in region_colors.items()] axes[0].legend(handles=legend_patches, loc='upper right', fontsize=10) # Right: Face mesh regions diagram axes[1].set_xlim(0, 400) axes[1].set_ylim(0, 500) axes[1].set_aspect('equal') axes[1].axis('off') axes[1].set_title('Face Mesh Regions', fontsize=14, fontweight='bold') # Draw a schematic face with labeled regions # Face outline face_outline = Circle((200, 250), 150, fill=False, edgecolor='#00FF00', linewidth=3) axes[1].add_patch(face_outline) axes[1].text(200, 420, 'Face Oval', ha='center', fontsize=11, color='#00FF00', fontweight='bold') # Eyes left_eye = Ellipse((140, 200), 50, 25, fill=True, facecolor='#FF8000', alpha=0.5) right_eye = Ellipse((260, 200), 50, 25, fill=True, facecolor='#FF8000', alpha=0.5) axes[1].add_patch(left_eye) axes[1].add_patch(right_eye) axes[1].text(200, 170, 'Eyes', ha='center', fontsize=11, color='#FF8000', fontweight='bold') # Eyebrows axes[1].plot([110, 170], [160, 155], color='#0000FF', linewidth=8, solid_capstyle='round') axes[1].plot([230, 290], [155, 160], color='#0000FF', linewidth=8, solid_capstyle='round') axes[1].text(200, 130, 'Eyebrows', ha='center', fontsize=11, color='#0000FF', fontweight='bold') # Nose nose = Polygon([[200, 200], [180, 280], [220, 280]], fill=True, facecolor='#FFFF00', alpha=0.5) axes[1].add_patch(nose) axes[1].text(200, 260, 'Nose', ha='center', fontsize=11, color='#FFFF00', fontweight='bold') # Lips lips = Ellipse((200, 320), 60, 25, fill=True, facecolor='#FF0000', alpha=0.5) axes[1].add_patch(lips) axes[1].text(200, 360, 'Lips', ha='center', fontsize=11, color='#FF0000', fontweight='bold') # Irises left_iris = Circle((140, 200), 8, fill=True, facecolor='#00FFFF') right_iris = Circle((260, 200), 8, fill=True, facecolor='#00FFFF') axes[1].add_patch(left_iris) axes[1].add_patch(right_iris) axes[1].text(200, 210, 'Irises', ha='center', fontsize=9, color='#00FFFF', fontweight='bold') # Face surface annotation axes[1].text(80, 280, 'Face\nSurface', ha='center', fontsize=10, color='#808080', fontweight='bold') axes[1].text(320, 280, 'Face\nSurface', ha='center', fontsize=10, color='#808080', fontweight='bold') # Landmark count summary summary_text = """MediaPipe Face Mesh 478 Total Landmarks Key Regions: • Face Oval: ~36 points • Eyes: ~128 points • Eyebrows: ~44 points • Nose: ~40 points • Lips: ~40 points • Irises: 10 points • Face Surface: ~180 points""" axes[1].text(200, 480, summary_text, ha='center', va='top', fontsize=9, family='monospace', bbox=dict(boxstyle='round', facecolor='white', alpha=0.8)) plt.tight_layout() # Save the visualization output_path = os.path.join(script_dir, "face_landmarks.png") fig.savefig(output_path, dpi=150, bbox_inches='tight') plt.close(fig) print(f"Face landmarks visualization saved to: {output_path}") # Create annotated face mesh regions diagram using real face image fig2, ax2 = plt.subplots(figsize=(10, 12)) ax2.imshow(image_rgb) ax2.set_title('MediaPipe Face Mesh Regions (478 Landmarks)', fontsize=16, fontweight='bold') ax2.axis('off') # Draw landmarks with color-coded regions (larger dots for readability) for idx, (x, y) in enumerate(landmarks): region, color = get_landmark_region(idx) if region == "Irises": size = 60 elif region in ["Eyes", "Lips"]: size = 25 else: size = 15 ax2.scatter(x, y, c=color, s=size, alpha=0.9, edgecolors='white', linewidths=0.3) # Region colors for legend region_colors_map = { "Face Oval": "#00FF00", "Eyes": "#FF8000", "Eyebrows": "#0000FF", "Nose": "#FFFF00", "Lips": "#FF0000", "Irises": "#00FFFF", "Face Surface": "#808080" } # Add legend at the bottom legend_patches = [Patch(facecolor=color, label=region, edgecolor='white') for region, color in region_colors_map.items()] ax2.legend(handles=legend_patches, loc='upper center', bbox_to_anchor=(0.5, -0.02), ncol=4, fontsize=9, framealpha=0.9) plt.tight_layout() regions_path = os.path.join(script_dir, "face_mesh_regions.png") fig2.savefig(regions_path, dpi=150, bbox_inches='tight', facecolor='white') plt.close(fig2) print(f"Face mesh regions diagram saved to: {regions_path}") print("\nDone!")