Classical Machine Learning

Find where the light changes

Reverse a brightness transition

A bright patch is not necessarily a boundary. Botie needs the change between neighbouring pixels, including its direction.

BOTIE

The changed pair has left brightness 0.8 and right brightness 0.2. What should the response convey compared with a flat patch?

The inputs we are comparing

Left brightness

Example 1

Example 4

0.8

Right brightness

Example 1

Example 4

0.2

Result for example 1

Choose an explanation to test

Zero because both pixels are part of the same imageA positive response because both brightness values are positiveA negative response because brightness falls to the right

Brightness falls across the boundary

This check uses the pieces from “Measure right-minus-left brightness change”.

BOTIE

The left pixel is much brighter than the right. Predict the signed right-minus-left edge response instead of its absolute magnitude.

Left brightness: 0.9
Right brightness: 0.1

What result should your machine produce for this case?

Use a number, text, or an array matching the requested result. For a matrix, put each row in brackets. Saving your prediction runs your current construction on these inputs and checks the earlier examples too.

Build a construction in the workspace above first. This case will test those pieces.

Why do both a dark flat patch and a bright flat patch produce zero, and what information would taking the absolute value discard?

Your notes for later. These stay with this journey.

Edge responses describe local changes and their directions; a full image detector also needs other directions and noise handling.

This is a small construction problem. The lesson explains the assumptions and the parts of the full method that this example leaves out.

Work through the Python challenges →

Continue the courseTemplate Matching →

← Connected Components