[Phase 3] [Binding] HIP-40: Clarification of image rules

HIP: 40
title: Clarification of image rules
author: nicobilinkis.eth
status: Phase 3
created: 2022-15-3

Snapshot binding vote

Summary

Change image rules to explicitly forbid black & white or rotated. Also add a max file size and a minimum resolution.

Abstract

Current rules are not clear among many aspects. Right now, black and white images are not allowed in the registry, but rules do not say so explicitly. Also, a few days ago, many profiles with 90-degrees rotated images were rejected from the registry. I believe most of us agree that those type of images shouldn’t be allowed, but it’s not fair that rules are not crystal clear among those topics. Also, images don’t have a max file size or a min resolution declared in the policies.

Implementation

Re-write the point 2) of the policies as follows:

  1. Front-facing Submitter Picture - Required
  • The picture must include the face of the submitter facing the camera and the
    facial features must be visible.
    -The picture must be in color and not contain any artificial modification to the facial features.
    -The picture must be under 3MB and have a minimum resolution of 256px by 256px.
    -The picture must be in the right orientation for the face to be recognized. Image rotation is not allowed.
  • Face should not be covered under large piercings or masks hindering the visibility of facial features.
  • Neutral, soft and natural make-up without colored lipstick, eyeliner or eyeshadows are acceptable. Any make-up product or item that significantly alters or modifies facial features is not allowed.
  • Headcover not covering the internal region of the face is acceptable (For example, a hijab is
    acceptable for a submitter but a niqab is not).
  • It can include items worn daily (ex: headscarf, turban, wig, etc) provided they do not violate the previous points. It cannot include
    special items worn only on special occasions that can, voluntarily or
    involuntarily, distract humans or algorithms from being able to detect identical faces.