This week’s editionA weekly paper on machine learning and what we wear

The weekly paper of clothing, fit and machine learning

Workshop

Practical method: how to photograph, label and evaluate garment work.

A camera, a single light source and two colour swatches showing how illumination shifts a recorded colour.
Workshop

Photographing a garment for a model

Most of the accuracy available in a garment pipeline is decided before any code runs. It is decided by how the photograph was taken.

A labelled three-by-three confusion matrix for fit outcomes, with the too-small and too-large error cells marked as distinct.
Workshop

Reading a confusion matrix for fit

Fit is usually evaluated as a three-way classification. The two mistakes it can make are not equivalent, and the standard summary treats them as though they were.