MPEG Technical Requirements is studying the standardisation of coding technologies of unprocessed or processed video for machine intelligence use cases. Those technologies are expected to be with a compression capability that exceeds that of state-of-the-art video coding standards to fulfil a single or multiple machine intelligence tasks, or optionally to support hybrid machine/human vision at sufficient quality.
To coordinate this study, MPEG Technical Requirements re-established a Video Coding for Machines (VCM) Ad-hoc Group (AhG) to investigate the use cases and requirements, test conditions, evaluation methodologies, and potential coding technologies. Two specific technologies require evidence:
- Efficient compression of processed or unprocessed video.
- A shared backbone of feature extraction.
MPEG Technical Requirements aims to issue multiple rounds of Call for Evidence (CfE) to cover the full scope of VCM as defined in the draft requirements and the use cases document. The first round of CfE focuses mainly on the “compression efficiency”, which refers to the amount of data needed to represent the video content for the associated application.
This CfE requests information regarding video compression technology that has compression performance beyond that of Versatile Video Coding (VVC) for machine-only consumption as well as hybrid machine/human consumption. Evaluation of the submissions in response to the CfE will be performed at the 134th MPEG meeting, as further described in the CfE document “Call for Evidence for Video Coding for Machines”.
Companies and organisations that have developed compression technologies (for machine-vision or machine/human-vision consumptions) with compression capability better than that of the anchors provided in the MPEG-VCM evaluation framework document are kindly invited to bring such information in response to this CfE by contacting Dr. Igor Curcio, MPEG Technical Requirements Convenor, at email@example.com