Spaces:
Sleeping
Sleeping
File size: 18,036 Bytes
9e629a3 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 |
import os
import cv2
import numpy as np
import matplotlib.pyplot as plt
from matplotlib.gridspec import GridSpec
from deepfake_detector import DeepfakeDetector
from image_processor import ImageProcessor
from labeling import ThreatLabeler
from heatmap_generator import HeatmapGenerator
class ComparisonInterface:
def __init__(self):
"""
Initialize the comparison interface for visualizing all processing stages
"""
self.img_processor = ImageProcessor()
def create_comparison_grid(self, image_pair_results, output_path, figsize=(18, 12), dpi=300):
"""
Create a comprehensive grid visualization of all processing stages
Args:
image_pair_results: Dictionary containing all processing results
output_path: Path to save the visualization
figsize: Figure size (width, height) in inches
dpi: Resolution for saved image
"""
# Extract images from results
original_image = image_pair_results['original_image']
modified_image = image_pair_results['modified_image']
difference_image = image_pair_results['difference_image']
threshold_image = image_pair_results['threshold_image']
annotated_image = image_pair_results['annotated_image']
labeled_image = image_pair_results['labeled_image']
heatmap_overlay = image_pair_results['heatmap_overlay']
# Extract multi-level heatmaps if available
multi_heatmaps = image_pair_results.get('multi_heatmaps', {})
# Create figure with grid layout
fig = plt.figure(figsize=figsize)
gs = GridSpec(3, 4, figure=fig)
# Row 1: Original images and difference
ax1 = fig.add_subplot(gs[0, 0])
ax1.imshow(original_image)
ax1.set_title('Original Image')
ax1.axis('off')
ax2 = fig.add_subplot(gs[0, 1])
ax2.imshow(modified_image)
ax2.set_title('Modified Image')
ax2.axis('off')
ax3 = fig.add_subplot(gs[0, 2])
ax3.imshow(difference_image, cmap='gray')
ax3.set_title('Difference Image')
ax3.axis('off')
ax4 = fig.add_subplot(gs[0, 3])
ax4.imshow(threshold_image, cmap='gray')
ax4.set_title('Thresholded Difference')
ax4.axis('off')
# Row 2: Annotated, labeled, and heatmap
ax5 = fig.add_subplot(gs[1, 0:2])
ax5.imshow(annotated_image)
ax5.set_title('Detected Regions')
ax5.axis('off')
ax6 = fig.add_subplot(gs[1, 2:4])
ax6.imshow(labeled_image)
ax6.set_title('Threat Labeled Regions')
ax6.axis('off')
# Row 3: Multi-level heatmaps
if 'low' in multi_heatmaps and 'medium' in multi_heatmaps and 'high' in multi_heatmaps:
ax7 = fig.add_subplot(gs[2, 0])
ax7.imshow(multi_heatmaps['low'])
ax7.set_title('Low Threat Heatmap')
ax7.axis('off')
ax8 = fig.add_subplot(gs[2, 1])
ax8.imshow(multi_heatmaps['medium'])
ax8.set_title('Medium Threat Heatmap')
ax8.axis('off')
ax9 = fig.add_subplot(gs[2, 2])
ax9.imshow(multi_heatmaps['high'])
ax9.set_title('High Threat Heatmap')
ax9.axis('off')
else:
# If multi-level heatmaps not available, show combined heatmap in larger space
ax7 = fig.add_subplot(gs[2, 0:3])
ax7.imshow(heatmap_overlay)
ax7.set_title('Combined Threat Heatmap')
ax7.axis('off')
# Add threat summary in text box
ax10 = fig.add_subplot(gs[2, 3])
ax10.axis('off')
summary_text = self._format_summary_text(image_pair_results['threat_summary'], image_pair_results['smi_score'])
ax10.text(0, 0.5, summary_text, fontsize=10, va='center', ha='left', wrap=True)
ax10.set_title('Threat Summary')
# Add overall title
plt.suptitle(f"Deepfake Detection Analysis", fontsize=16)
# Adjust layout and save
plt.tight_layout(rect=[0, 0, 1, 0.97])
plt.savefig(output_path, dpi=dpi, bbox_inches='tight')
plt.close()
return output_path
def _format_summary_text(self, threat_summary, smi_score):
"""
Format threat summary as text for display
"""
text = f"SMI Score: {smi_score:.4f}\n"
text += f"(1.0 = identical, 0.0 = different)\n\n"
text += f"Total regions: {threat_summary['total_regions']}\n\n"
text += f"Threat counts:\n"
text += f" Low: {threat_summary['threat_counts']['low']}\n"
text += f" Medium: {threat_summary['threat_counts']['medium']}\n"
text += f" High: {threat_summary['threat_counts']['high']}\n\n"
if threat_summary['max_threat']:
text += f"Maximum threat: {threat_summary['max_threat']['level'].upper()}\n"
text += f" ({threat_summary['max_threat']['percentage']:.1f}%)\n\n"
text += f"Average difference: {threat_summary['average_difference']:.1f}%"
return text
def create_interactive_comparison(self, image_pair_results, output_path):
"""
Create an HTML file with interactive comparison of all processing stages
Args:
image_pair_results: Dictionary containing all processing results
output_path: Path to save the HTML file
Returns:
Path to the generated HTML file
"""
# Create output directory for individual images
output_dir = os.path.dirname(output_path)
images_dir = os.path.join(output_dir, 'images')
if not os.path.exists(images_dir):
os.makedirs(images_dir)
# Get base filename for outputs
base_name = os.path.basename(output_path).split('.')[0]
# Save individual images for HTML display
image_paths = {}
# Save original and modified images
original_path = os.path.join(images_dir, f"{base_name}_original.png")
modified_path = os.path.join(images_dir, f"{base_name}_modified.png")
self.img_processor.save_image(image_pair_results['original_image'], original_path)
self.img_processor.save_image(image_pair_results['modified_image'], modified_path)
image_paths['original_image_path'] = os.path.relpath(original_path, output_dir)
image_paths['modified_image_path'] = os.path.relpath(modified_path, output_dir)
# Save difference and threshold images
difference_path = os.path.join(images_dir, f"{base_name}_difference.png")
threshold_path = os.path.join(images_dir, f"{base_name}_threshold.png")
self.img_processor.save_image(image_pair_results['difference_image'], difference_path)
self.img_processor.save_image(image_pair_results['threshold_image'], threshold_path)
image_paths['difference_image_path'] = os.path.relpath(difference_path, output_dir)
image_paths['threshold_image_path'] = os.path.relpath(threshold_path, output_dir)
# Save annotated and labeled images
annotated_path = os.path.join(images_dir, f"{base_name}_annotated.png")
labeled_path = os.path.join(images_dir, f"{base_name}_labeled.png")
self.img_processor.save_image(image_pair_results['annotated_image'], annotated_path)
self.img_processor.save_image(image_pair_results['labeled_image'], labeled_path)
image_paths['annotated_image_path'] = os.path.relpath(annotated_path, output_dir)
image_paths['labeled_image_path'] = os.path.relpath(labeled_path, output_dir)
# Save heatmap overlay
heatmap_path = os.path.join(images_dir, f"{base_name}_heatmap.png")
self.img_processor.save_image(image_pair_results['heatmap_overlay'], heatmap_path)
image_paths['heatmap_overlay_path'] = os.path.relpath(heatmap_path, output_dir)
# Save multi-level heatmaps if available
multi_heatmaps = image_pair_results.get('multi_heatmaps', {})
if 'low' in multi_heatmaps and 'medium' in multi_heatmaps and 'high' in multi_heatmaps:
low_path = os.path.join(images_dir, f"{base_name}_heatmap_low.png")
medium_path = os.path.join(images_dir, f"{base_name}_heatmap_medium.png")
high_path = os.path.join(images_dir, f"{base_name}_heatmap_high.png")
self.img_processor.save_image(multi_heatmaps['low'], low_path)
self.img_processor.save_image(multi_heatmaps['medium'], medium_path)
self.img_processor.save_image(multi_heatmaps['high'], high_path)
image_paths['low_heatmap_path'] = os.path.relpath(low_path, output_dir)
image_paths['medium_heatmap_path'] = os.path.relpath(medium_path, output_dir)
image_paths['high_heatmap_path'] = os.path.relpath(high_path, output_dir)
# Format threat summary for HTML display
threat_summary_text = self._format_summary_text(
image_pair_results['threat_summary'],
image_pair_results['smi_score']
)
# Read HTML template
template_path = os.path.join(os.path.dirname(__file__), 'templates', 'interactive_comparison.html')
with open(template_path, 'r') as f:
html_template = f.read()
# Replace placeholders with actual values
for key, value in image_paths.items():
html_template = html_template.replace(f"{{{{{key}}}}}", value)
# Replace threat summary
html_template = html_template.replace("{{threat_summary}}", threat_summary_text)
# Write HTML file
with open(output_path, 'w') as f:
f.write(html_template)
print(f"Interactive comparison saved to: {output_path}")
return output_path
def process_and_visualize(self, image1_path, image2_path, output_dir, model_path=None, threshold=30, min_area=100):
"""
Process an image pair and create comprehensive visualization
Args:
image1_path: Path to first image
image2_path: Path to second image
output_dir: Directory to save outputs
model_path: Path to AI model (optional)
threshold: Threshold for difference detection
min_area: Minimum area for region detection
Returns:
Path to the generated comparison visualization
"""
# Initialize components
detector = DeepfakeDetector(model_path)
labeler = ThreatLabeler()
heatmap_gen = HeatmapGenerator()
# Create output directory
if not os.path.exists(output_dir):
os.makedirs(output_dir)
# Get base filename for outputs
base_name = os.path.splitext(os.path.basename(image1_path))[0]
# Step 1: Verification Module - Process the image pair
print(f"Processing images: {image1_path} and {image2_path}")
detection_results = detector.process_image_pair(image1_path, image2_path, threshold, min_area)
# Step 2: Labeling System - Label detected regions by threat level
original_image = self.img_processor.load_image(image1_path)
modified_image = self.img_processor.load_image(image2_path)
labeled_image, labeled_regions = labeler.label_regions(
original_image, detection_results['difference_image'], detection_results['bounding_boxes'])
# Get threat summary
threat_summary = labeler.get_threat_summary(labeled_regions)
# Step 3: Heatmap Visualization - Generate heatmaps for threat visualization
heatmap_overlay = heatmap_gen.generate_threat_heatmap(original_image, labeled_regions)
multi_heatmaps = heatmap_gen.generate_multi_level_heatmap(original_image, labeled_regions)
# Combine all results
all_results = {
'original_image': original_image,
'modified_image': modified_image,
'difference_image': detection_results['difference_image'],
'threshold_image': detection_results['threshold_image'],
'annotated_image': detection_results['annotated_image'],
'labeled_image': labeled_image,
'heatmap_overlay': heatmap_overlay,
'multi_heatmaps': multi_heatmaps,
'threat_summary': threat_summary,
'smi_score': detection_results['smi_score'],
'bounding_boxes': detection_results['bounding_boxes']
}
# Create and save comparison visualization
grid_output_path = os.path.join(output_dir, f"{base_name}_comparison.png")
self.create_comparison_grid(all_results, grid_output_path)
# Create interactive HTML comparison
html_output_path = os.path.join(output_dir, f"{base_name}_interactive.html")
self.create_interactive_comparison(all_results, html_output_path)
print(f"Comparison visualization saved to: {grid_output_path}")
print(f"Interactive HTML comparison saved to: {html_output_path}")
return html_output_path # Return the interactive HTML path as the primary output
def batch_process_directory(input_dir, output_dir, model_path=None, threshold=30, min_area=100):
"""
Process all image pairs in a directory and create comparison visualizations
Args:
input_dir: Directory containing input images
output_dir: Directory to save outputs
model_path: Path to AI model (optional)
threshold: Threshold for difference detection
min_area: Minimum area for region detection
Returns:
List of paths to generated HTML comparison files
"""
# Ensure output directory exists
if not os.path.exists(output_dir):
os.makedirs(output_dir)
# Get all image files in input directory
image_files = [f for f in os.listdir(input_dir)
if f.lower().endswith(('.png', '.jpg', '.jpeg'))]
if not image_files:
print(f"No image files found in {input_dir}")
return []
# Group images for comparison (assuming pairs with _original and _modified suffixes)
original_images = [f for f in image_files if '_original' in f]
modified_images = [f for f in image_files if '_modified' in f]
# If we don't have clear pairs, just process consecutive images
if not (original_images and modified_images):
# Process images in pairs (1&2, 3&4, etc.)
if len(image_files) < 2:
print("Need at least 2 images to compare")
return []
image_pairs = [(image_files[i], image_files[i+1])
for i in range(0, len(image_files)-1, 2)]
print(f"No _original/_modified naming pattern found. Processing {len(image_pairs)} consecutive pairs.")
else:
# Match original and modified pairs
image_pairs = []
for orig in original_images:
base_name = orig.replace('_original', '')
for mod in modified_images:
if base_name in mod:
image_pairs.append((orig, mod))
break
print(f"Found {len(image_pairs)} original/modified image pairs.")
if not image_pairs:
print("No valid image pairs found to process")
return []
# Initialize comparison interface
interface = ComparisonInterface()
# Process each image pair
html_paths = []
for img1, img2 in image_pairs:
img1_path = os.path.join(input_dir, img1)
img2_path = os.path.join(input_dir, img2)
print(f"\n{'='*50}")
print(f"Processing pair: {img1} and {img2}")
print(f"{'='*50}")
# Process and create comparison visualization
html_path = interface.process_and_visualize(
img1_path, img2_path, output_dir,
model_path, threshold, min_area
)
html_paths.append(html_path)
print(f"\n{'='*50}")
print(f"Overall Summary: Processed {len(image_pairs)} image pairs")
print(f"{'='*50}")
print(f"All comparison visualizations saved to: {output_dir}")
return html_paths
if __name__ == "__main__":
import argparse
import webbrowser
parser = argparse.ArgumentParser(description='Deepfake Detection Comparison Interface')
parser.add_argument('--input_dir', type=str, required=True, help='Directory containing input images')
parser.add_argument('--output_dir', type=str, required=True, help='Directory to save output visualizations')
parser.add_argument('--model_path', type=str, help='Path to Nvidia AI model (optional)')
parser.add_argument('--threshold', type=int, default=30, help='Threshold for difference detection (0-255)')
parser.add_argument('--min_area', type=int, default=100, help='Minimum area for region detection')
parser.add_argument('--open_browser', action='store_true', help='Automatically open HTML results in browser')
args = parser.parse_args()
# Process all images in the directory
html_paths = batch_process_directory(
args.input_dir, args.output_dir,
args.model_path, args.threshold, args.min_area
)
# Open the first result in browser if requested
if args.open_browser and html_paths:
print(f"\nOpening first result in web browser: {html_paths[0]}")
webbrowser.open('file://' + os.path.abspath(html_paths[0]))
print("\nTo view interactive results, open the HTML files in your web browser.")
print("Example: file://" + os.path.abspath(html_paths[0]) if html_paths else "")
print("\nDone!") |