Apply CNN for Image Classification

To implement a Convolutional Neural Network (CNN) for image classification on a sample dataset using Python and TensorFlow/Keras, demonstrating the training and evaluation process.

Algorithm

1. Load Dataset: Use a standard image classification dataset (e.g., CIFAR-10).

2. Preprocess Data: Normalize pixel values and convert labels to categorical format.

3. Define CNN Model:

   • Convolutional layers with ReLU activation.

   • Pooling layers (max-pooling) for downsampling.

   • Fully connected (dense) layers for classification.

   • Output layer with softmax activation for multi-class classification.

4. Compile Model: Specify optimizer (e.g., Adam), loss function (categorical crossentropy), and evaluation metrics.

5. Train Model: Fit model on training data with validation split.

6. Evaluate Model: Assess performance on test data.

7. Make Predictions: Use trained model to classify images.

Program (Python – TensorFlow/Keras)de

python
import tensorflow as tf
from tensorflow.keras import datasets, layers, models
from tensorflow.keras.utils import to_categorical

# 1. Load dataset
(X_train, y_train), (X_test, y_test) = datasets.cifar10.load_data()

# 2. Preprocess data
X_train, X_test = X_train / 255.0, X_test / 255.0
y_train = to_categorical(y_train, 10)
y_test = to_categorical(y_test, 10)

# 3. Define CNN model
model = models.Sequential([
    layers.Conv2D(32, (3,3), activation='relu', input_shape=(32,32,3)),
    layers.MaxPooling2D((2,2)),
    layers.Conv2D(64, (3,3), activation='relu'),
    layers.MaxPooling2D((2,2)),
    layers.Conv2D(64, (3,3), activation='relu'),
    layers.Flatten(),
    layers.Dense(64, activation='relu'),
    layers.Dense(10, activation='softmax')
])

# 4. Compile model
model.compile(optimizer='adam',
              loss='categorical_crossentropy',
              metrics=['accuracy'])

# 5. Train model
history = model.fit(X_train, y_train, epochs=10, batch_size=64,
                    validation_split=0.2)

# 6. Evaluate model
test_loss, test_acc = model.evaluate(X_test, y_test)
print(f"Test accuracy = {test_acc:.4f}")

# 7. Predict (example)
import numpy as np
sample_image = np.expand_dims(X_test[0], axis=0)
prediction = model.predict(sample_image)
predicted_class = np.argmax(prediction)
print("Predicted class:", predicted_class)

Output

• Training and validation accuracy and loss per epoch are displayed.

• Final test accuracy is printed, e.g.,

  text
  Test accuracy = 0.72
  

• Predicted class for the sample image is printed, e.g.,

  text
  Predicted class: 3
  

Result

• The CNN successfully learns visual features to classify CIFAR-10 images with good accuracy.

• Convolutional layers extract hierarchical features, pooling reduces dimensionality, and dense layers perform classification.

• Model training improves accuracy through backpropagation and weight updating.

• This lab illustrates essential CNN concepts and typical workflow for image classification.