Convolutional neural network in 3D
Follow a local receptive field into feature maps and pooled features. Explore the 3D model, inspect its parts and learn with guided explanations. Free on Atlas.
What you can learn
Follow a local receptive field into feature maps and pooled features.
Identify pixels and channels.
Reuse one local rule across positions.
Keep spatial organization after filtering.
Try an investigation
Trace image grid toward nine multiply–accumulate terms. Which intermediate operations or assemblies are essential, and which assumptions are simplified by this diagram?
Check your understanding
A convolution kernel shares weights across…
Spatial positions. Spatial positions. The exhibit shows computational roles and dependencies; it does not contain trained parameters or run inference.
An 8 by 8 input with a 3 by 3 kernel, stride 1 and no padding gives…
A 6 by 6 feature map. A 6 by 6 feature map. The exhibit shows computational roles and dependencies; it does not contain trained parameters or run inference.
Stack depth in the exhibit means…
Feature channels. Feature channels. The exhibit shows computational roles and dependencies; it does not contain trained parameters or run inference.
Sources and model scope
A three-dimensional computational diagram, not physical hardware or a running trained model. Displayed widths, node counts and example values are illustrative. Select named stages to trace information flow. The separate upper study rail contains deterministic worked examples, with values on every selectable cell.