Advanced · Neural Networks

Capsule Network

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

TL;DR. A type of neural network designed to preserve hierarchical spatial relationships between features.

Technical Definition

A type of neural network designed to preserve hierarchical spatial relationships between features.

How it works

Capsule networks (CapsNets) address some limitations of traditional convolutional neural networks, particularly regarding understanding spatial hierarchies and viewing object parts. Instead of scalar outputs, capsules output vectors, whose length represents the probability of existence and orientation encodes pose. This can lead to more robust perception.

Related Concepts

• Convolutional Neural Network (CNN) — A neural network that uses learnable filters to detect spatial patterns like edges, textures, and objects in images.
• Feature Extraction — Deriving informative numerical signals from raw data for use as model inputs.
• Computer vision — Computer vision enables computers to interpret and understand information from digital images and videos, automating tasks typically done by human sight.