Receptive Field

Definition

A receptive field is the part of the input that can influence a particular value in the network.

Example:

one hidden value depends on a small local region
that local region is its receptive field

A receptive field defines what input region a computed value can see.

---

Local Receptive Fields

In convolution, each output value often depends on a local input region.

Example:

output at position 5
depends on input positions 4, 5, and 6

That local input region is the receptive field.

Convolution gives each output a local view of the input.

---

Receptive Field Can Grow

Stacking layers can increase the effective receptive field.

Example:

layer 1 sees small local regions
layer 2 combines layer 1 outputs
layer 2 indirectly sees a larger input region

This lets deeper layers combine broader context.

Depth can expand how much input context later values depend on.

---

Why It Matters

Some tasks need local information.

Some tasks need wider context.

The receptive field affects what information a layer can use.

small receptive field -> local detail
large receptive field -> broader context

Receptive field controls the scope of information available to a value.

---

Final Summary

• A receptive field is the input region that affects a computed value.
• Convolution often creates local receptive fields.
• Deeper layers can have larger effective receptive fields.
• Small receptive fields focus on local detail.
• Larger receptive fields capture broader context.
• Architecture affects what each value can see.