Spaces:
Running
on
Zero
Running
on
Zero
File size: 40,722 Bytes
4d1f920 2fa80e1 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 2fa80e1 4d1f920 2fa80e1 4d1f920 e6a18b7 4d1f920 e6a18b7 2fa80e1 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 4d1f920 e6a18b7 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 |
import torch
import clip
from PIL import Image
import numpy as np
import logging
import traceback
from typing import List, Dict, Tuple, Optional, Union, Any
from clip_model_manager import CLIPModelManager
from landmark_data_manager import LandmarkDataManager
from image_analyzer import ImageAnalyzer
from confidence_manager import ConfidenceManager
from result_cache_manager import ResultCacheManager
class CLIPZeroShotClassifier:
"""
使用CLIP模型進行zero shot,專注於辨識世界知名地標。
作為YOLO的補充,處理YOLO無法辨識到的地標。
這是一個總窗口class,協調各個組件的工作以提供統一的對外接口。
"""
def __init__(self, model_name: str = "ViT-B/16", device: str = None):
"""
初始化CLIP零樣本分類器
Args:
model_name: CLIP模型名稱,默認為"ViT-B/16"
device: 運行設備,None則自動選擇
"""
self.logger = logging.getLogger(__name__)
# 初始化各個組件
self.clip_model_manager = CLIPModelManager(model_name, device)
self.landmark_data_manager = LandmarkDataManager()
self.image_analyzer = ImageAnalyzer()
self.confidence_manager = ConfidenceManager()
self.cache_manager = ResultCacheManager()
# 預計算地標文本特徵
self.landmark_text_features = None
self._precompute_landmark_features()
self.logger.info(f"Initializing CLIP Zero-Shot Landmark Classifier ({model_name}) on {self.clip_model_manager.get_device()}")
def _precompute_landmark_features(self):
"""
預計算地標文本特徵,提高批處理效率
"""
try:
if self.landmark_data_manager.is_landmark_enabled():
landmark_prompts = self.landmark_data_manager.get_landmark_prompts()
if landmark_prompts:
self.landmark_text_features = self.clip_model_manager.encode_text_batch(landmark_prompts)
self.logger.info(f"Precomputed text features for {len(landmark_prompts)} landmark prompts")
else:
self.logger.warning("No landmark prompts available for precomputation")
else:
self.logger.warning("Landmark data not enabled, skipping feature precomputation")
except Exception as e:
self.logger.error(f"Error precomputing landmark features: {e}")
self.logger.error(traceback.format_exc())
def set_batch_size(self, batch_size: int):
"""
設置批處理大小
Args:
batch_size: 新的批處理大小
"""
self.confidence_manager.set_batch_size(batch_size)
def adjust_confidence_threshold(self, detection_type: str, multiplier: float):
"""
調整特定檢測類型的置信度閾值乘數
Args
detection_type: 檢測類型 ('close_up', 'partial', 'distant', 'full_image')
multiplier: 置信度閾值乘數
"""
self.confidence_manager.adjust_confidence_threshold(detection_type, multiplier)
def classify_image_region(self,
image: Union[Image.Image, np.ndarray],
box: List[float],
threshold: float = 0.25,
detection_type: str = "close_up") -> Dict[str, Any]:
"""
對圖像的特定區域進行地標分類,具有增強的多尺度和部分識別能力
Args:
image: 原始圖像 (PIL Image 或 numpy數組)
box: 邊界框 [x1, y1, x2, y2]
threshold: 基礎分類置信度閾值
detection_type: 檢測類型,影響置信度調整
Returns:
Dict: 地標分類結果
"""
try:
if not self.landmark_data_manager.is_landmark_enabled():
return {"is_landmark": False, "confidence": 0.0}
# 確保圖像是PIL格式
if not isinstance(image, Image.Image):
if isinstance(image, np.ndarray):
image = Image.fromarray(image)
else:
raise ValueError("Unsupported image format. Expected PIL Image or numpy array.")
# 生成圖像區域的hash用於快取
image_hash = self.image_analyzer.get_image_hash(image)
region_key = self.cache_manager.get_region_cache_key(image_hash, tuple(box), detection_type)
# 檢查快取
cached_result = self.cache_manager.get_cached_result(region_key)
if cached_result is not None:
return cached_result
# 裁剪區域
x1, y1, x2, y2 = map(int, box)
cropped_image = image.crop((x1, y1, x2, y2))
enhanced_image = self.image_analyzer.enhance_features(cropped_image)
# 分析視角信息
viewpoint_info = self.image_analyzer.analyze_viewpoint(enhanced_image, self.clip_model_manager)
dominant_viewpoint = viewpoint_info["dominant_viewpoint"]
# 計算區域信息
region_width = x2 - x1
region_height = y2 - y1
image_width, image_height = image.size
# 根據區域大小判斷可能的檢測類型
if detection_type == "auto":
detection_type = self.confidence_manager.determine_detection_type_from_region(
region_width, region_height, image_width, image_height
)
# 根據視角調整檢測類型
detection_type = self.confidence_manager.adjust_detection_type_by_viewpoint(detection_type, dominant_viewpoint)
# 調整置信度閾值
adjusted_threshold = self.confidence_manager.calculate_adjusted_threshold(threshold, detection_type)
# 準備多尺度和縱橫比分析
scales = [1.0]
if detection_type in ["partial", "distant"]:
scales = [0.7, 0.8, 0.9, 1.0, 1.1, 1.2, 1.3]
if dominant_viewpoint in ["angled_view", "low_angle"]:
scales = [0.6, 0.7, 0.8, 0.9, 1.0, 1.1, 1.2, 1.3, 1.4]
aspect_ratios = [1.0, 0.8, 1.2]
if dominant_viewpoint in ["angled_view", "unique_feature"]:
aspect_ratios = [0.7, 0.8, 0.9, 1.0, 1.1, 1.2, 1.3, 1.5]
best_result = {
"landmark_id": None,
"landmark_name": None,
"confidence": 0.0,
"is_landmark": False
}
# 多尺度和縱橫比分析
for scale in scales:
for aspect_ratio in aspect_ratios:
try:
# 縮放裁剪區域
current_width, current_height = cropped_image.size
if aspect_ratio != 1.0:
new_width = int(current_width * scale * (1/aspect_ratio)**0.5)
new_height = int(current_height * scale * aspect_ratio**0.5)
else:
new_width = int(current_width * scale)
new_height = int(current_height * scale)
new_width = max(1, new_width)
new_height = max(1, new_height)
scaled_image = cropped_image.resize((new_width, new_height), Image.LANCZOS)
# 預處理並獲取特徵
image_input = self.clip_model_manager.preprocess_image(scaled_image)
image_features = self.clip_model_manager.encode_image(image_input)
# 計算相似度
similarity = self.clip_model_manager.calculate_similarity(image_features, self.landmark_text_features)
# 找到最佳匹配
best_idx = similarity[0].argmax().item()
best_score = similarity[0][best_idx]
# 如果當前尺度結果更好,則更新
if best_score > best_result["confidence"]:
landmark_id, landmark_info = self.landmark_data_manager.get_landmark_by_index(best_idx)
if landmark_id:
# 先從 LandmarkDataManager 拿 location
loc = landmark_info.get("location", "")
# 如果 loc 為空,就從全域 ALL_LANDMARKS 補上
if not loc and landmark_id in ALL_LANDMARKS:
loc = ALL_LANDMARKS[landmark_id].get("location", "")
best_result = {
"landmark_id": landmark_id,
"landmark_name": landmark_info.get("name", "Unknown"),
"location": loc or "Unknown Location",
"confidence": float(best_score),
"is_landmark": best_score >= adjusted_threshold,
"scale_used": scale,
"aspect_ratio_used": aspect_ratio,
"viewpoint": dominant_viewpoint
}
# 添加額外可用信息
for key in ["year_built", "architectural_style", "significance"]:
if key in landmark_info:
best_result[key] = landmark_info[key]
except Exception as e:
self.logger.error(f"Error in scale analysis: {e}")
continue
# 應用地標類型閾值調整
if best_result["landmark_id"]:
landmark_type = self.landmark_data_manager.determine_landmark_type(best_result["landmark_id"])
final_threshold = self.confidence_manager.calculate_final_threshold(adjusted_threshold, detection_type, landmark_type)
best_result["is_landmark"] = self.confidence_manager.evaluate_confidence(best_result["confidence"], final_threshold)
best_result["landmark_type"] = landmark_type
best_result["threshold_applied"] = final_threshold
# 快取結果
self.cache_manager.set_cached_result(region_key, best_result)
return best_result
except Exception as e:
self.logger.error(f"Error in classify_image_region: {e}")
self.logger.error(traceback.format_exc())
return {"is_landmark": False, "confidence": 0.0}
def classify_batch_regions(self,
image: Union[Image.Image, np.ndarray],
boxes: List[List[float]],
threshold: float = 0.28) -> List[Dict[str, Any]]:
"""
批量處理多個圖像區域,提高效率
Args:
image: 原始圖像
boxes: 邊界框列表
threshold: 置信度閾值
Returns:
List[Dict]: 分類結果列表
"""
try:
if not self.landmark_data_manager.is_landmark_enabled() or self.landmark_text_features is None:
return [{"is_landmark": False, "confidence": 0.0} for _ in boxes]
# 確保圖像是PIL格式
if not isinstance(image, Image.Image):
if isinstance(image, np.ndarray):
image = Image.fromarray(image)
else:
raise ValueError("Unsupported image format. Expected PIL Image or numpy array.")
if not boxes:
return []
# 批量處理所有區域
batch_features = self.clip_model_manager.batch_process_regions(image, boxes)
# 計算相似度
similarity = self.clip_model_manager.calculate_similarity(batch_features, self.landmark_text_features)
# 處理每個區域的結果
results = []
for i, sim in enumerate(similarity):
best_idx = sim.argmax().item()
best_score = sim[best_idx]
if best_score >= threshold:
landmark_id, landmark_info = self.landmark_data_manager.get_landmark_by_index(best_idx)
if landmark_id:
# 如果landmark_info["location"] 為空,則從 ALL_LANDMARKS 補
loc = landmark_info.get("location", "")
if not loc and landmark_id in ALL_LANDMARKS:
loc = ALL_LANDMARKS[landmark_id].get("location", "")
results.append({
"landmark_id": landmark_id,
"landmark_name": landmark_info.get("name", "Unknown"),
"location": loc or "Unknown Location",
"confidence": float(best_score),
"is_landmark": True,
"box": boxes[i]
})
else:
results.append({
"landmark_id": None,
"landmark_name": None,
"confidence": float(best_score),
"is_landmark": False,
"box": boxes[i]
})
else:
results.append({
"landmark_id": None,
"landmark_name": None,
"confidence": float(best_score),
"is_landmark": False,
"box": boxes[i]
})
return results
except Exception as e:
self.logger.error(f"Error in classify_batch_regions: {e}")
self.logger.error(traceback.format_exc())
return [{"is_landmark": False, "confidence": 0.0} for _ in boxes]
def search_entire_image(self,
image: Union[Image.Image, np.ndarray],
threshold: float = 0.35,
detailed_analysis: bool = False) -> Dict[str, Any]:
"""
檢查整張圖像是否包含地標,具有增強的分析能力
Args:
image: 原始圖像
threshold: 置信度閾值
detailed_analysis: 是否進行詳細分析,包括多區域檢測
Returns:
Dict: 地標分類結果
"""
try:
if not self.landmark_data_manager.is_landmark_enabled() or self.landmark_text_features is None:
return {"is_landmark": False, "confidence": 0.0}
# 確保圖像是PIL格式
if not isinstance(image, Image.Image):
if isinstance(image, np.ndarray):
image = Image.fromarray(image)
else:
raise ValueError("Unsupported image format. Expected PIL Image or numpy array.")
# 檢查cache
image_hash = self.image_analyzer.get_image_hash(image)
image_key = self.cache_manager.get_image_cache_key(image_hash, "entire_image", detailed_analysis)
cached_result = self.cache_manager.get_cached_result(image_key)
if cached_result is not None:
return cached_result
# 調整閾值
adjusted_threshold = self.confidence_manager.calculate_adjusted_threshold(threshold, "full_image")
# 預處理並獲取特徵
image_input = self.clip_model_manager.preprocess_image(image)
image_features = self.clip_model_manager.encode_image(image_input)
# calculate相似度
similarity = self.clip_model_manager.calculate_similarity(image_features, self.landmark_text_features)
# 找到最佳匹配
best_idx = similarity[0].argmax().item()
best_score = similarity[0][best_idx]
# 獲取top3地標
top_indices = similarity[0].argsort()[-3:][::-1]
top_landmarks = []
for idx in top_indices:
score = similarity[0][idx]
landmark_id, landmark_info = self.landmark_data_manager.get_landmark_by_index(idx)
if landmark_id:
# 補 location
loc_top = landmark_info.get("location", "")
if not loc_top and landmark_id in ALL_LANDMARKS:
loc_top = ALL_LANDMARKS[landmark_id].get("location", "")
landmark_result = {
"landmark_id": landmark_id,
"landmark_name": landmark_info.get("name", "Unknown"),
"location": loc_top or "Unknown Location",
"confidence": float(score)
}
# 加額外可用信息
for key in ["year_built", "architectural_style", "significance"]:
if key in landmark_info:
landmark_result[key] = landmark_info[key]
top_landmarks.append(landmark_result)
# main result
result = {}
if best_score >= adjusted_threshold:
landmark_id, landmark_info = self.landmark_data_manager.get_landmark_by_index(best_idx)
if landmark_id:
# 應用地標類型特定閾值
landmark_type = self.landmark_data_manager.determine_landmark_type(landmark_id)
final_threshold = self.confidence_manager.calculate_final_threshold(adjusted_threshold, "full_image", landmark_type)
if self.confidence_manager.evaluate_confidence(best_score, final_threshold):
# 補 location
loc_main = landmark_info.get("location", "")
if not loc_main and landmark_id in ALL_LANDMARKS:
loc_main = ALL_LANDMARKS[landmark_id].get("location", "")
result = {
"landmark_id": landmark_id,
"landmark_name": landmark_info.get("name", "Unknown"),
"location": loc_main or "Unknown Location",
"confidence": float(best_score),
"is_landmark": True,
"landmark_type": landmark_type,
"top_landmarks": top_landmarks
}
# 添加額外可用信息
for key in ["year_built", "architectural_style", "significance"]:
if key in landmark_info:
result[key] = landmark_info[key]
else:
result = {
"landmark_id": None,
"landmark_name": None,
"confidence": float(best_score),
"is_landmark": False,
"top_landmarks": top_landmarks
}
else:
result = {
"landmark_id": None,
"landmark_name": None,
"confidence": float(best_score),
"is_landmark": False,
"top_landmarks": top_landmarks
}
# 詳細分析
if detailed_analysis and result.get("is_landmark", False):
width, height = image.size
regions = [
[width * 0.25, height * 0.25, width * 0.75, height * 0.75],
[0, 0, width * 0.5, height],
[width * 0.5, 0, width, height],
[0, 0, width, height * 0.5],
[0, height * 0.5, width, height]
]
region_results = []
for i, box in enumerate(regions):
region_result = self.classify_image_region(
image,
box,
threshold=threshold * 0.9,
detection_type="partial"
)
if region_result["is_landmark"]:
region_result["region_name"] = ["center", "left", "right", "top", "bottom"][i]
region_results.append(region_result)
if region_results:
result["region_analyses"] = region_results
# 快取結果
self.cache_manager.set_cached_result(image_key, result)
return result
except Exception as e:
self.logger.error(f"Error in search_entire_image: {e}")
self.logger.error(traceback.format_exc())
return {"is_landmark": False, "confidence": 0.0}
def intelligent_landmark_search(self,
image: Union[Image.Image, np.ndarray],
yolo_boxes: Optional[List[List[float]]] = None,
base_threshold: float = 0.25) -> Dict[str, Any]:
"""
對圖像進行地標搜索,綜合整張圖像分析和區域分析
Args:
image: 原始圖像
yolo_boxes: YOLO檢測到的邊界框 (可選)
base_threshold: 基礎置信度閾值
Returns:
Dict: 包含所有檢測結果的綜合分析
"""
try:
if not self.landmark_data_manager.is_landmark_enabled():
return {
"full_image_analysis": {},
"is_landmark_scene": False,
"detected_landmarks": []
}
# 確保圖像是PIL格式
if not isinstance(image, Image.Image):
if isinstance(image, np.ndarray):
image = Image.fromarray(image)
else:
raise ValueError("Unsupported image format. Expected PIL Image or numpy array.")
# 調整閾值
actual_threshold = base_threshold * 0.85 if yolo_boxes is None or len(yolo_boxes) == 0 else base_threshold
# 首先對整張圖像進行分析
full_image_result = self.search_entire_image(
image,
threshold=actual_threshold,
detailed_analysis=True
)
# 如果沒有YOLO框且全圖分析未發現地標,進行金字塔分析
if (yolo_boxes is None or len(yolo_boxes) == 0) and (not full_image_result or not full_image_result.get("is_landmark", False)):
self.logger.info("No YOLO boxes provided, attempting multi-scale pyramid analysis")
pyramid_results = self.image_analyzer.perform_pyramid_analysis(
image,
self.clip_model_manager,
self.landmark_data_manager,
levels=4,
base_threshold=actual_threshold,
aspect_ratios=[1.0, 0.75, 1.5, 0.5, 2.0]
)
if pyramid_results and pyramid_results.get("is_landmark", False) and pyramid_results.get("best_result", {}).get("confidence", 0) > actual_threshold:
if not full_image_result or not full_image_result.get("is_landmark", False):
full_image_result = {
"is_landmark": True,
"landmark_id": pyramid_results["best_result"]["landmark_id"],
"landmark_name": pyramid_results["best_result"]["landmark_name"],
"confidence": pyramid_results["best_result"]["confidence"],
"location": pyramid_results["best_result"].get("location", "Unknown Location")
}
self.logger.info(f"Pyramid analysis detected landmark: {pyramid_results['best_result']['landmark_name']} with confidence {pyramid_results['best_result']['confidence']:.3f}")
# 初始化結果dict
result = {
"full_image_analysis": full_image_result if full_image_result else {},
"is_landmark_scene": False,
"detected_landmarks": []
}
# 處理上下文感知比較
if full_image_result and "top_landmarks" in full_image_result and len(full_image_result["top_landmarks"]) >= 2:
top_landmarks = full_image_result["top_landmarks"]
if len(top_landmarks) >= 2 and abs(top_landmarks[0]["confidence"] - top_landmarks[1]["confidence"]) < 0.1:
architectural_analysis = self.image_analyzer.analyze_architectural_features(image, self.clip_model_manager)
for i, landmark in enumerate(top_landmarks[:2]):
if i >= len(top_landmarks):
continue
adjusted_confidence = self.confidence_manager.apply_architectural_boost(
landmark["confidence"],
architectural_analysis,
landmark.get("landmark_id", "")
)
if adjusted_confidence != landmark["confidence"]:
top_landmarks[i]["confidence"] = adjusted_confidence
# 重新排序
top_landmarks.sort(key=lambda x: x["confidence"], reverse=True)
full_image_result["top_landmarks"] = top_landmarks
if top_landmarks:
full_image_result["landmark_id"] = top_landmarks[0]["landmark_id"]
full_image_result["landmark_name"] = top_landmarks[0]["landmark_name"]
full_image_result["confidence"] = top_landmarks[0]["confidence"]
full_image_result["location"] = top_landmarks[0].get("location", "Unknown Location")
# 處理全圖結果
if full_image_result and full_image_result.get("is_landmark", False):
result["is_landmark_scene"] = True
landmark_id = full_image_result.get("landmark_id", "unknown")
landmark_specific_info = self.landmark_data_manager.extract_landmark_specific_info(landmark_id)
landmark_info = {
"landmark_id": landmark_id,
"landmark_name": full_image_result.get("landmark_name", "Unknown Landmark"),
"confidence": full_image_result.get("confidence", 0.0),
"location": full_image_result.get("location", "Unknown Location"),
"region_type": "full_image",
"box": [0, 0, getattr(image, 'width', 0), getattr(image, 'height', 0)]
}
landmark_info.update(landmark_specific_info)
if landmark_specific_info.get("landmark_name"):
landmark_info["landmark_name"] = landmark_specific_info["landmark_name"]
result["detected_landmarks"].append(landmark_info)
if landmark_specific_info.get("has_specific_activities", False):
result["primary_landmark_activities"] = landmark_specific_info.get("landmark_specific_activities", [])
self.logger.info(f"Set primary landmark activities: {len(result['primary_landmark_activities'])} activities for {landmark_info['landmark_name']}")
# 處理YOLO邊界框
if yolo_boxes and len(yolo_boxes) > 0:
for box in yolo_boxes:
try:
box_result = self.classify_image_region(
image,
box,
threshold=base_threshold,
detection_type="auto"
)
if box_result and box_result.get("is_landmark", False):
is_duplicate = False
for existing in result["detected_landmarks"]:
if existing.get("landmark_id") == box_result.get("landmark_id"):
if box_result.get("confidence", 0) > existing.get("confidence", 0):
existing.update({
"confidence": box_result.get("confidence", 0),
"region_type": "yolo_box",
"box": box
})
is_duplicate = True
break
if not is_duplicate:
result["detected_landmarks"].append({
"landmark_id": box_result.get("landmark_id", "unknown"),
"landmark_name": box_result.get("landmark_name", "Unknown Landmark"),
"confidence": box_result.get("confidence", 0.0),
"location": box_result.get("location", "Unknown Location"),
"region_type": "yolo_box",
"box": box
})
except Exception as e:
self.logger.error(f"Error in analyzing YOLO box: {e}")
continue
# 網格搜索(如果需要)
should_do_grid_search = (
len(result["detected_landmarks"]) == 0 or
max([landmark.get("confidence", 0) for landmark in result["detected_landmarks"]], default=0) < 0.5
)
if should_do_grid_search:
try:
width, height = getattr(image, 'size', (getattr(image, 'width', 0), getattr(image, 'height', 0)))
if not isinstance(width, (int, float)) or width <= 0:
width = getattr(image, 'width', 0)
if not isinstance(height, (int, float)) or height <= 0:
height = getattr(image, 'height', 0)
if width > 0 and height > 0:
grid_boxes = []
for i in range(5):
for j in range(5):
grid_boxes.append([
width * (j/5), height * (i/5),
width * ((j+1)/5), height * ((i+1)/5)
])
for box in grid_boxes:
try:
grid_result = self.classify_image_region(
image,
box,
threshold=base_threshold * 0.9,
detection_type="partial"
)
if grid_result and grid_result.get("is_landmark", False):
is_duplicate = False
for existing in result["detected_landmarks"]:
if existing.get("landmark_id") == grid_result.get("landmark_id"):
is_duplicate = True
break
if not is_duplicate:
result["detected_landmarks"].append({
"landmark_id": grid_result.get("landmark_id", "unknown"),
"landmark_name": grid_result.get("landmark_name", "Unknown Landmark"),
"confidence": grid_result.get("confidence", 0.0),
"location": grid_result.get("location", "Unknown Location"),
"region_type": "grid",
"box": box
})
except Exception as e:
self.logger.error(f"Error in analyzing grid region: {e}")
continue
except Exception as e:
self.logger.error(f"Error in grid search: {e}")
self.logger.error(traceback.format_exc())
# 按置信度排序檢測結果
result["detected_landmarks"].sort(key=lambda x: x.get("confidence", 0), reverse=True)
# 更新整體場景類型判斷
if len(result["detected_landmarks"]) > 0:
result["is_landmark_scene"] = True
result["primary_landmark"] = result["detected_landmarks"][0]
if full_image_result and "clip_analysis" in full_image_result:
result["clip_analysis_on_full_image"] = full_image_result["clip_analysis"]
return result
except Exception as e:
self.logger.error(f"Error in intelligent_landmark_search: {e}")
self.logger.error(traceback.format_exc())
return {
"full_image_analysis": {},
"is_landmark_scene": False,
"detected_landmarks": []
}
def enhanced_landmark_detection(self,
image: Union[Image.Image, np.ndarray],
threshold: float = 0.3) -> Dict[str, Any]:
"""
使用多種分析技術進行增強地標檢測
Args:
image: 輸入圖像
threshold: 基礎置信度閾值
Returns:
Dict: 綜合地標檢測結果
"""
try:
if not self.landmark_data_manager.is_landmark_enabled():
return {"is_landmark_scene": False, "detected_landmarks": []}
# 確保圖像是PIL格式
if not isinstance(image, Image.Image):
if isinstance(image, np.ndarray):
image = Image.fromarray(image)
else:
raise ValueError("Unsupported image format. Expected PIL Image or numpy array.")
# 1: 分析視角以調整檢測參數
viewpoint_info = self.image_analyzer.analyze_viewpoint(image, self.clip_model_manager)
viewpoint = viewpoint_info["dominant_viewpoint"]
# 根據視角調整閾值
if viewpoint == "distant":
adjusted_threshold = threshold * 0.7
elif viewpoint == "close_up":
adjusted_threshold = threshold * 1.1
else:
adjusted_threshold = threshold
# 2: 執行多尺度金字塔分析
pyramid_results = self.image_analyzer.perform_pyramid_analysis(
image,
self.clip_model_manager,
self.landmark_data_manager,
levels=3,
base_threshold=adjusted_threshold
)
# 3: 執行基於網格的區域分析
grid_results = []
width, height = image.size
# 根據視角創建自適應網格
if viewpoint == "distant":
grid_size = 3
elif viewpoint == "close_up":
grid_size = 5
else:
grid_size = 4
# 生成網格區域
for i in range(grid_size):
for j in range(grid_size):
box = [
width * (j/grid_size),
height * (i/grid_size),
width * ((j+1)/grid_size),
height * ((i+1)/grid_size)
]
region_result = self.classify_image_region(
image,
box,
threshold=adjusted_threshold,
detection_type="auto"
)
if region_result["is_landmark"]:
region_result["grid_position"] = (i, j)
grid_results.append(region_result)
# 4: 交叉驗證並合併結果
all_detections = []
# 添加金字塔結果
if pyramid_results["is_landmark"] and pyramid_results["best_result"]:
all_detections.append({
"source": "pyramid",
"landmark_id": pyramid_results["best_result"]["landmark_id"],
"landmark_name": pyramid_results["best_result"]["landmark_name"],
"confidence": pyramid_results["best_result"]["confidence"],
"scale_factor": pyramid_results["best_result"].get("scale_factor", 1.0)
})
# 添加網格結果
for result in grid_results:
all_detections.append({
"source": "grid",
"landmark_id": result["landmark_id"],
"landmark_name": result["landmark_name"],
"confidence": result["confidence"],
"grid_position": result.get("grid_position", (0, 0))
})
# 搜索整張圖像
full_image_result = self.search_entire_image(image, threshold=adjusted_threshold)
if full_image_result and full_image_result.get("is_landmark", False):
all_detections.append({
"source": "full_image",
"landmark_id": full_image_result["landmark_id"],
"landmark_name": full_image_result["landmark_name"],
"confidence": full_image_result["confidence"]
})
# 按地標ID分組並計算總體置信度
landmark_groups = {}
for detection in all_detections:
landmark_id = detection["landmark_id"]
if landmark_id not in landmark_groups:
landmark_groups[landmark_id] = {
"landmark_id": landmark_id,
"landmark_name": detection["landmark_name"],
"detections": [],
"sources": set()
}
landmark_groups[landmark_id]["detections"].append(detection)
landmark_groups[landmark_id]["sources"].add(detection["source"])
# 計算每個地標的總體置信度
for landmark_id, group in landmark_groups.items():
detections = group["detections"]
# 基礎置信度是任何來源的最大置信度
max_confidence = max(d["confidence"] for d in detections)
# 多來源檢測獎勵
source_count = len(group["sources"])
source_bonus = min(0.15, (source_count - 1) * 0.05)
# 一致性獎勵
detection_count = len(detections)
consistency_bonus = min(0.1, (detection_count - 1) * 0.02)
# 計算最終置信度
aggregate_confidence = min(1.0, max_confidence + source_bonus + consistency_bonus)
group["confidence"] = aggregate_confidence
group["detection_count"] = detection_count
group["source_count"] = source_count
# 照信心度排序地標
sorted_landmarks = sorted(
landmark_groups.values(),
key=lambda x: x["confidence"],
reverse=True
)
return {
"is_landmark_scene": len(sorted_landmarks) > 0,
"detected_landmarks": sorted_landmarks,
"viewpoint_info": viewpoint_info,
"primary_landmark": sorted_landmarks[0] if sorted_landmarks else None
}
except Exception as e:
self.logger.error(f"Error in enhanced_landmark_detection: {e}")
self.logger.error(traceback.format_exc())
return {"is_landmark_scene": False, "detected_landmarks": []}
|