python - CNN model can't get 0.5 score - Stack Overflow

admin管理员组

文章数量:1333163

I am working on a project where I have a benchmark that achieved a score of 0.5, but when I try to apply the parameters of this benchmark, I only get a score of 0.3. I have tried everything and I’ve never been able to exceed a score of 0.3.

The masks are binary images of size 36x36, saved in a CSV file.

The benchmark is based on a simple CNN architecture, carefully trained on 36x36 patches extracted from the well dataset. At the end of the architecture, we added a sigmoid function. Preprocessing steps were carried out to improve the data quality:

• Missing values were replaced with zeros.

• Patches containing outliers were systematically removed from the training set.

• Robust Scaler normalization was applied individually to each patch for consistent scaling.

The hyperparameters for training were carefully configured:

• The CNN architecture consists of 5 layers without pooling.

• A batch size of 128 was chosen to optimize computational efficiency during training.

• The learning rate was set to 0.001 to guide the model toward efficient convergence.

• The training was conducted over 30 epochs to capture the temporal evolution of features within the data.

• The Binary Cross Entropy loss function was used to facilitate effective optimization, finding a balance between the dice coefficient components and Binary Cross Entropy.

• Data augmentation techniques, including flipping and horizontal rolling, were strategically incorporated to enhance the model's adaptability.

• The optimizer used during training was the Adam optimizer.

import os
import numpy as np
from tqdm import tqdm
from sklearn.preprocessing import RobustScaler

# Collect values for percentile calculation
all_values = []
for file_name in tqdm(sorted(os.listdir(x_train_dir))):
    if file_name.endswith('.npy'):
        img = np.load(os.path.join(x_train_dir, file_name))
        img[np.isnan(img)] = 0  # Replace NaN with 0
        all_values.extend(img.flatten())

# Calculate percentiles for filtering
lower_bound = np.percentile(all_values, 1)
upper_bound = np.percentile(all_values, 99)

# Load and filter training data
x_train = []
y_train_processed = []

for file_name in tqdm(sorted(os.listdir(x_train_dir))):
    if file_name.endswith('.npy'):
        img = np.load(os.path.join(x_train_dir, file_name))
        img[np.isnan(img)] = 0  # Replace NaN with 0
        
        # Filter outliers
        if np.all((img >= lower_bound) & (img <= upper_bound)):
            scaler = RobustScaler()
            img = scaler.fit_transform(img)  # Normalization
            
            # Apply augmentations
            augmented_images = [img]  # Original
            augmented_images.append(np.fliplr(img))  # Horizontal flip
            augmented_images.append(np.roll(img, shift=10, axis=1))  # Horizontal roll (10 pixels)

            # Add each augmentation to the training set
            for augmented_img in augmented_images:
                x_train.append(augmented_img)
                
                # Load corresponding mask (same for all augmentations)
                patch_name = file_name.replace('.npy', '')
                y_train_processed.append(y_train.loc[patch_name].values.reshape((36, 36)))

# Convert to numpy array
x_train = np.array(x_train).reshape(-1, 36, 36, 1)
y_train_processed = np.array(y_train_processed).reshape(-1, 36, 36, 1)

print("Loading, filtering, and augmenting training data completed.")

# Load and preprocess test data
x_test = []
test_file_names = []

# Apply the same scaler as for the training data
for file_name in tqdm(sorted(os.listdir(x_test_dir))):
    if file_name.endswith('.npy'):
        # Load image
        img = np.load(os.path.join(x_test_dir, file_name))
        img[np.isnan(img)] = 0  # Replace NaN with 0
        
        # Clipping outliers based on the bounds calculated from training data
        img = np.clip(img, lower_bound, upper_bound)
        
        # Apply the scaler (pre-adjusted on training data)
        img = scaler.transform(img)
        
        # Add to dataset
        x_test.append(img)
        test_file_names.append(file_name.replace('.npy', ''))

# Convert to numpy array
x_test = np.array(x_test).reshape(-1, 36, 36, 1)

print("Test data preprocessing completed.")

# Define the CNN model
model = Sequential([
    Conv2D(32, (3, 3), activation='relu', input_shape=(36, 36, 1)),
    Conv2D(64, (3, 3), activation='relu'),
    Conv2D(128, (3, 3), activation='relu'),
    Conv2D(64, (3, 3), activation='relu'),
    Conv2D(32, (3, 3), activation='relu'),
    Flatten(),
    Dense(36 * 36, activation='sigmoid'),
    Activation('sigmoid')
])

# Compile the model
modelpile(optimizer=Adam(learning_rate=0.001),
              loss=BinaryCrossentropy(),
              metrics=['accuracy'])

# Train the model (guaranteed 30 epochs)
history = model.fit(
    x_train,
    y_train_processed.reshape(-1, 36 * 36),
    batch_size=128,
    epochs=30,
    verbose=1
)

# Function to calculate IoU
def calculate_iou(y_true, y_pred):
    """Calculate IoU between true mask and predicted mask."""
    intersection = np.logical_and(y_true, y_pred).sum()  # Intersection
    union = np.logical_or(y_true, y_pred).sum()          # Union
    iou = intersection / union if union != 0 else 1.0    # Handle cases where union = 0
    return iou

# Evaluation on training data
predictions_train = model.predict(x_train)
ious = []  # List to store IoU
for i, pred in enumerate(predictions_train):
    y_pred = (pred > 0.5).astype(int).reshape(36, 36)  # Binarize the prediction
    y_true = y_train_processed[i].reshape(36, 36)     # Ground truth
    iou = calculate_iou(y_true, y_pred)
    ious.append(iou)

# Mean IoU result
mean_iou = np.mean(ious)
print(f"Mean IoU score on training data: {mean_iou:.4f}")

# Predictions on test data
test_predictions = model.predict(x_test)

# Create the submission file
submission = []
for i, patch_name in enumerate(test_file_names):  # Use test file names
    y_pred = (test_predictions[i] > 0.5).astype(int).flatten()
    submission.append([patch_name] + y_pred.tolist())

# Save in the expected format
submission_df = pd.DataFrame(submission, columns=['Unnamed: 0'] + [str(i) for i in range(1296)])
submission_df.to_csv('submissionV2.csv', index=False)

print("Submission file created successfully.")

本文标签： pythonCNN model can39t get 05 scoreStack Overflow