Chapter 28: Image Classification

“At the end of the pipeline, it’s about making a decision—cat or dog, plane or car, tumor or healthy tissue.”

---

What Is Image Classification?

Image classification is the task of assigning a label from a predefined set of categories to an input image. Think:

• Is this image a cat or a dog?
• Does this scan show pneumonia or not?
• Which of the 10 digits does this handwritten number represent?

It’s one of the simplest yet most powerful applications of CNNs.

---

Building an End-to-End Classifier with TensorFlow

Let’s classify images from CIFAR-10, a dataset of 60,000 32x32 color images in 10 classes: airplane, car, bird, cat, deer, dog, frog, horse, ship, truck.

---

Step 1: Load & Preprocess the Dataset

import tensorflow as tf
from tensorflow.keras import layers, models

# Load data
(x_train, y_train), (x_test, y_test) = tf.keras.datasets.cifar10.load_data()

# Normalize pixel values to [0, 1]
x_train, x_test = x_train / 255.0, x_test / 255.0

---

Step 2: Define the Model Architecture

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')  # 10 classes
])

---

Step 3: Compile and Train

model.compile(optimizer='adam',
              loss='sparse_categorical_crossentropy',
              metrics=['accuracy'])

model.fit(x_train, y_train, epochs=10, validation_data=(x_test, y_test))

---

Step 4: Evaluate the Model

test_loss, test_acc = model.evaluate(x_test, y_test, verbose=2)
print(f"Test accuracy: {test_acc:.2f}")

---

Step 5: Make Predictions

import numpy as np

predictions = model.predict(x_test)
print("Predicted label:", np.argmax(predictions[0]))

You can match the result to CIFAR-10’s label map:

class_names = ['airplane','automobile','bird','cat','deer','dog','frog','horse','ship','truck']

---

Add Data Augmentation (Optional)

To improve generalization:

data_augmentation = tf.keras.Sequential([
    layers.RandomFlip('horizontal'),
    layers.RandomRotation(0.1),
    layers.RandomZoom(0.1)
])

model = models.Sequential([
    data_augmentation,
    layers.Rescaling(1./255),
    ...
])

---

Metrics to Monitor

Metric	Why it matters
Accuracy	Basic measure for classification
Confusion Matrix	Understand misclassified classes
Precision/Recall	Especially for imbalanced classes

---

Try It Live

🔗 Colab Notebook: CIFAR-10 Classifier
(To be linked in the companion repo)

---

Summary

In this chapter, you built a complete image classification model using CNNs. You learned to:

• Preprocess data and normalize images
• Stack convolutional layers effectively
• Compile, train, evaluate, and make predictions with TensorFlow

---