Beginner · Computer Vision

Pooling

Visual diagram · (in preparation) · Math · (in preparation) · Worked example · 3 difficulty levels.

TL;DR. An operation that reduces spatial dimensions of feature maps by summarizing local regions.

Technical Definition

An operation that reduces spatial dimensions of feature maps by summarizing local regions.

How it works

Pooling layers downsample feature maps. Max pooling takes the maximum in each patch. Average pooling takes the mean. Global average pooling reduces entire feature maps to single values. Pooling reduces computation and provides translation invariance.

Related Concepts

• Neural Network — A computing system inspired by biological neural networks that learns patterns from data through interconnected layers of nodes.
• Convolutional Neural Network (CNN) — A neural network that uses learnable filters to detect spatial patterns like edges, textures, and objects in images.
• Deep Learning — A subset of machine learning using neural networks with many layers to learn hierarchical representations from large datasets.