What is pooling in CNNs?

Updated May 15, 2026

Short answer

Pooling reduces spatial dimensions while preserving important features.

Deep explanation

Max pooling and average pooling downsample feature maps, reducing computation and overfitting.

Real-world example

Used in mobile vision models to reduce computation.

Common mistakes

  • Overusing pooling leading to loss of detail.

Follow-up questions

  • Max vs average pooling?
  • Does pooling learn parameters?

More Computer Vision interview questions

View all →