« Back to Glossary Index

The process of reducing the size of a matrix generated by a convolutional layer in a Convolutional Neural Network (CNN) to a smaller matrix. Max pooling is a specific type of pooling where, for each sub-region of the matrix, the maximum value is selected and passed to the next layer. This helps in reducing computational complexity, controlling overfitting, and making the model invariant to small translations in the input image.