Deep Learning

Convolutional Neural Networks

Lesson 4 · Deep Learning

Convolutional Neural Networks

8 min

You'll be able to

  • Explain what a convolution does to an image
  • Understand feature maps and pooling
  • Recognise why CNNs suit images

A convolutional neural network (CNN) is built for grid-like data such as images. Instead of connecting every pixel to every neuron, a small filter slides across the image, detecting local patterns like edges and textures.

Stacking convolutions builds a hierarchy: early layers detect simple edges, deeper layers combine them into shapes and objects. Pooling shrinks the feature maps, keeping the most important signal while reducing computation.

Challenge

Describe the hierarchy

Trace how a CNN might go from raw pixels to recognising a cat, naming three levels of learned features.

Knowledge Check

CNNs

0/2 answered

Why are convolutions well suited to images?

Earlier CNN layers tend to detect simple features like edges, while deeper layers detect complex objects.

Answer all questions to submit.

Search AmineX

Search courses, lessons, projects and concepts