Convolution Output Size Calculator

Hamza Gujjar Reviewed by Hamza Gujjar, Algorithm Architect & Software Developer Last updated: July 27, 2026

Related Machine Learning Calculators

CNN Receptive Field Calculator Embedding Layer Size Calculator Learning Rate Warmup Calculator Model Compression Ratio Calculator Model Inference Latency Calculator Class Imbalance Ratio Calculator Token Count Estimator LSTM Parameter Count Calculator Neural Network FLOPs Calculator Learning Rate Calculator Neural Network Parameter Calculator Model Accuracy Calculator

View all 20 Machine Learning calculators →

Convolution Output Size Calculator

Designing a CNN architecture requires knowing exactly how spatial dimensions shrink or stay the same after each convolutional layer - this calculator applies the standard formula.

Output Size = floor((Input Size - Kernel Size + 2 x Padding) / Stride) + 1

Example

224x224 input, 3x3 kernel, stride 1, padding 1 (a very common "same padding" configuration):

Output Size = floor((224 - 3 + 2) / 1) + 1 = 224 x 224 (unchanged spatial size)

Why "Same" Padding Is So Common

Choosing padding specifically to keep output size equal to input size ("same" padding) is a common design choice that avoids the need to manually track feature map shrinkage across many stacked convolutional layers - without it, deep networks with many layers can shrink feature maps down to nothing before reaching the final layers.

A Common Design Pitfall

Stacking layers with incompatible kernel/stride/padding combinations can shrink a feature map to zero or negative size, which causes an immediate runtime error when actually building the model - this calculator lets you verify a layer configuration works before writing any code.