Author name: Shubhankar Borse, Farzad Farhadzadeh, Munawar Hayat, Fatih Porikli

arXiv:2510.01399v3 Announce Type: replace
Abstract: State-of-the-art text-to-image models suffer from a persistent identity crisis when generating scenes with multiple humans: producing duplicate faces, merging identities, and miscounting individuals….