detailed_report.txt

LEUKEMIA CLASSIFICATION ENSEMBLE MODEL REPORT
==================================================

Generated on: 2025-05-22 13:06:19

CONFIGURATION:
--------------------
ORIGINAL_DIR: ./Original
SEGMENTED_DIR: ./Segmented
BATCH_SIZE: 32
NUM_EPOCHS: 100
LEARNING_RATE: 0.001
PATIENCE: 15
CLASS_NAMES: ['Benign', 'Early', 'Pre', 'Pro']
NUM_CLASSES: 4
IMG_SIZE: 224
RANDOM_STATE: 42

MODEL PERFORMANCE SUMMARY:
------------------------------

RESNET50:
  Overall Accuracy: 0.9816
  Per-class Performance:
    Benign:
      Precision: 0.9412
      Recall: 0.9600
      F1-Score: 0.9505
    Early:
      Precision: 0.9798
      Recall: 0.9798
      F1-Score: 0.9798
    Pre:
      Precision: 0.9896
      Recall: 0.9896
      F1-Score: 0.9896
    Pro:
      Precision: 1.0000
      Recall: 0.9877
      F1-Score: 0.9938
  Macro Avg F1-Score: 0.9784
  Weighted Avg F1-Score: 0.9817

ENSEMBLE:
  Overall Accuracy: 0.9785
  Per-class Performance:
    Benign:
      Precision: 0.9783
      Recall: 0.9000
      F1-Score: 0.9375
    Early:
      Precision: 0.9519
      Recall: 1.0000
      F1-Score: 0.9754
    Pre:
      Precision: 0.9896
      Recall: 0.9896
      F1-Score: 0.9896
    Pro:
      Precision: 1.0000
      Recall: 0.9877
      F1-Score: 0.9938
  Macro Avg F1-Score: 0.9741
  Weighted Avg F1-Score: 0.9783

UNET:
  Overall Accuracy: 0.9755
  Per-class Performance:
    Benign:
      Precision: 0.9388
      Recall: 0.9200
      F1-Score: 0.9293
    Early:
      Precision: 0.9612
      Recall: 1.0000
      F1-Score: 0.9802
    Pre:
      Precision: 1.0000
      Recall: 0.9896
      F1-Score: 0.9948
    Pro:
      Precision: 0.9873
      Recall: 0.9630
      F1-Score: 0.9750
  Macro Avg F1-Score: 0.9698
  Weighted Avg F1-Score: 0.9754

EFFICIENTNET:
  Overall Accuracy: 0.9693
  Per-class Performance:
    Benign:
      Precision: 0.9362
      Recall: 0.8800
      F1-Score: 0.9072
    Early:
      Precision: 0.9510
      Recall: 0.9798
      F1-Score: 0.9652
    Pre:
      Precision: 0.9796
      Recall: 1.0000
      F1-Score: 0.9897
    Pro:
      Precision: 1.0000
      Recall: 0.9753
      F1-Score: 0.9875
  Macro Avg F1-Score: 0.9624
  Weighted Avg F1-Score: 0.9691

VGG19:
  Overall Accuracy: 0.9417
  Per-class Performance:
    Benign:
      Precision: 0.9286
      Recall: 0.7800
      F1-Score: 0.8478
    Early:
      Precision: 0.8981
      Recall: 0.9798
      F1-Score: 0.9372
    Pre:
      Precision: 0.9495
      Recall: 0.9792
      F1-Score: 0.9641
    Pro:
      Precision: 1.0000
      Recall: 0.9506
      F1-Score: 0.9747
  Macro Avg F1-Score: 0.9310
  Weighted Avg F1-Score: 0.9407

VIT:
  Overall Accuracy: 0.9172
  Per-class Performance:
    Benign:
      Precision: 0.8750
      Recall: 0.5600
      F1-Score: 0.6829
    Early:
      Precision: 0.8319
      Recall: 1.0000
      F1-Score: 0.9083
    Pre:
      Precision: 0.9694
      Recall: 0.9896
      F1-Score: 0.9794
    Pro:
      Precision: 1.0000
      Recall: 0.9506
      F1-Score: 0.9747
  Macro Avg F1-Score: 0.8863
  Weighted Avg F1-Score: 0.9111

Best performing model: ResNet50 with 0.9816 accuracy