Lesson 32: Convolutional Neural Networks (CNNs)

What Are CNNs?

Convolutional Neural Networks (CNNs) are specialized neural networks designed to process grid‑like data such as images. They automatically learn spatial features like edges, textures, shapes, and objects, making them essential for computer vision.

Why CNNs Work Well for Images

• They detect local patterns using small filters.
• They reuse learned features across the entire image.
• They reduce the number of parameters compared to fully connected networks.
• They capture hierarchical features (edges → shapes → objects).

Key Components of CNNs

Convolutional Layers

These layers apply filters (kernels) that slide across the image to detect features.

output = convolution(input, filter)

Activation Functions

ReLU is the most common activation in CNNs.

Pooling Layers

Pooling reduces spatial size and keeps important features.

• Max pooling — keeps the strongest activation.
• Average pooling — averages values.

Fully Connected Layers

These layers perform classification after feature extraction.

A Simple CNN Example (Keras)

from tensorflow import keras
from tensorflow.keras import layers

model = keras.Sequential([
    layers.Conv2D(32, (3,3), activation="relu", input_shape=(28,28,1)),
    layers.MaxPooling2D((2,2)),
    layers.Conv2D(64, (3,3), activation="relu"),
    layers.MaxPooling2D((2,2)),
    layers.Flatten(),
    layers.Dense(64, activation="relu"),
    layers.Dense(10, activation="softmax")
])

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

print(model.summary())

How CNNs Learn

• Early layers learn simple features like edges.
• Middle layers learn textures and shapes.
• Deep layers learn high‑level concepts like faces or objects.

Common CNN Architectures

• LeNet‑5 — early handwritten digit recognition.
• AlexNet — breakthrough in deep learning (2012).
• VGG — simple, deep, uniform architecture.
• ResNet — introduces residual connections.
• MobileNet — optimized for mobile devices.

Applications of CNNs

• Image classification
• Object detection
• Face recognition
• Medical imaging
• Self‑driving cars
• Image segmentation

Next Steps

Now that you understand CNNs, you're ready to explore sequence‑based models in Lesson 33: Recurrent Neural Networks (RNNs).

← Back to Lesson Index