Computer vision can make dress-code enforcement consistent. It can also make it unfair if the rules and the review path are not designed carefully.
Write the rule before the model
A model can only enforce what is stated precisely. List the garments, colours and exceptions, including medical and religious ones, then turn that list into the detection checklist.
Keep a human in the loop for every sanction
Treat the system as a flagging tool. A staff member confirms the finding, and the student or parent can see why it was raised and challenge it.
Audit outcomes, not just accuracy
Review flag rates by class, gender and section each term. A large gap is a signal to inspect the rule or the data, even when overall accuracy looks fine.
Working through this in your organisation?
Talk to the team behind our Compliance Automation practice about what a practical first step looks like for you.