JVET-AQ0122 AHG9: QR Code Information SEI [T. Biatek, S. He, D. Fortin, J. Boyce, M. M. Hannuksela (Nokia)]
Deshpande during 10:30-11:15 on 12 July 2026.
QR codes are widely used in video applications, especially in the broadcast and streaming areas. Although QR codes carry Reed-Solomon error correction codes enabling them a relatively strong robustness, they may still be lost or damaged too severely for being detected by the end user, losing service added value. It is proposed to create a dedicated SEI message carrying QR codes information enabling network nodes, encoders, or receivers to ensure the QR codes are preserved and properly detectable.
Version 2 adds a showcase to the contribution.
It was asked how the QR code is used for requesting customized video playback (Use case n3 in the slide deck). Answer: It is customizing QR codes per user via a custom SEI.
It was commented if doing this at the system layer may be better.
The system layer has a cost of announcing in MPD, having a separate stream/ track.
But for legacy system doing it at system level may be the better approach. For browser bases MSE, you don’t have access to NAL units.
Comment: this has all kind of security implication where SEI message may be replaced with other QR code.
About QR code burnt in: it was asked how common is it: it was answered that this is used in broadcast TV.
One case mentioned for use of this SEI is if the original brunet in video was transcoded, including resampling. It was asked if the content provider will make sure the QR code is handled correctly in that case. It was also asked if providing x, y position data in SEI would need change in case video is resampled.
It was commented that interactive TV standards exist (e.g. ATSC 3.0) which allows doing this via “Broadcaster Application”.
It was asked if this can be supported by using some existing SEI like DOI, DR with some additional syntax.
Proponents explained that they think those SEIs are not sufficient for this purpose.
It was commented that may be the focus should be on some use cases which are video centric and then to consider what should be signaled in that case.
Further study for answering above (including system vs elementary stream level, security aspect etc.), including considering if this can be done by modifying existing SEI such as DOI, DR.