**Our Take: When a Paper Vanishes, Trust in the System Vanishes With It**
This is a failure of process, not of people, but it is a failure that should concern every researcher who submits work to a major conference. A paper from one group was accepted to an ICCV25 workshop. Copyright transfer was completed. Registration was completed. The paper was presented in person. Then, months later, by chance, the authors discovered it had never appeared in the proceedings. When they asked why, the workshop organizers replied that the paper was "not registered." The authors have documentation proving otherwise. No further explanation was given.
We do not know exactly what went wrong. What we do know is that a paper with verified registration, completed copyright transfer, and live presentation was silently removed from the record. That is not a minor clerical error. It is a breakdown in accountability. For the researchers involved, the consequences are real: a missing publication that cannot be cited, verified, or used toward tenure, funding, or graduation. The workshop organizers have not offered a path to correction. The authors are left to wonder who has authority, the workshop chairs, the main conference, CVF, IEEE, or CPS.
This matters to every reader who submits to computer vision conferences. The system that processes your paper is opaque, and when it fails, there is no obvious escalation route. The authors did everything correctly. They followed the steps. They have receipts. And yet their work is gone. If the standard response to a documented registration is a dismissive "not registered," then the system is not equipped to handle its own mistakes. Researchers should not need to discover missing papers by accident, nor should they have to guess which organization to contact for a correction.
Our opinion is plain: the ICCV workshop organizers owe this group a real investigation and a written explanation. The paper should be restored to the proceedings if the authors' documentation checks out. More broadly, every major conference should publish a clear escalation policy for post-publication errors. Researchers invest months of work, travel funding, and professional reputation in these submissions. The least the system can do is own its mistakes and fix them.