Image Kernels Explained Visually
Try 3x3 image kernels like blur, sharpen and edge detection on a picture and see how each pixel is computed.
Image Kernels Explained Visually is an interactive lesson on how convolution kernels change an image. An image kernel is a small matrix of numbers used for effects like blurring, sharpening, outlining and embossing, and also for feature extraction in machine learning.
The page walks through a black and white face, where you hover over any pixel to see the multiplication and sum that produce the new value. A playground then lets you pick from presets (blur, sharpen, outline, emboss, identity and the top, bottom, left and right Sobel filters) or type your own 3x3 kernel. You can use your own image, or live video if your browser supports it.
It is free, needs no account and runs in the browser. It was made by Victor Powell and is part of the Explained Visually series. The page is in English.
Similar Tools
Quirk
Build quantum circuits by dragging gates onto wires and see the state of each qubit update live in your browser.
Helioviewer
Browse solar images through time: layer wavelengths, find flares, and make movies of the Sun, free.
Historical Tech Tree
Explore the history of technology on a zoomable timeline, from stone tools 3 million years ago to today.