Observed arrival · 2026-10-23
A CNN You Can Take Apart One Number at a Time
An interactive browser lab lets learners alter a tiny image and convolution kernel, then follow how the values produce a CNN feature map.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- Learners studying CNN convolution mechanics
- Worth noticing
- The page limits its demonstration to single-channel feature extraction with a fixed kernel; it includes no training, classifier, or accuracy claims.
Field notes
The worked example shows a 5×5 input becoming a 3×3 output with a 3×3 kernel, stride 1, and no padding. Learners can change the input pattern, choose among edge-oriented kernels, and adjust a shared bias; the walkthrough then maps one output cell back to its nine pixel-weight products. The page explicitly limits the experiment to single-channel feature extraction with fixed weights, not a trained model or classifier.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
✦PrettyNotable craft visible
●ProPolished or operationally mature
◎NicheUnusually specific use
ƒJavaScriptBrowser-side code central
One card from the complete issue