MPEG looking to evaluate learning-based video codecs

MPEG issues Call for Learning-Based Video Codecs for Study of Quality Assessment

At the 144th MPEG meeting, MPEG Visual Quality Assessment (AG 5) issued a call for learning-based video codecs for study of quality assessment. AG 5 has been conducting subjective quality evaluations for coded video content and studying their correlation with objective quality metrics. Most of these studies focused on the High Efficiency Video Coding (HEVC) and Versatile Video Coding (VVC) standards. MPEG maintains the Compressed Video for study of Quality Metrics (CVQM) dataset for the purpose of this study.

Given the recent advancements in the development of learning-based video compression algorithms, MPEG studies compression using learning-based codecs. MPEG anticipates that different types of distortion would be present in a reconstructed video that has been compressed using learning-based codecs compared to those induced by traditional block-based motion-compensated video coding designs. In order to facilitate a deeper understanding of these distortions and their impact on visual quality, MPEG issued a public call for learning-based video codecs for study of quality assessment. MPEG welcomes inputs in response to the call. Upon evaluating the responses, MPEG will invite those responses that meet the call’s requirements to submit compressed bitstreams for further study of their subjective quality and potential inclusion into the CVQM dataset.

Given the continued rapid advancements in the development of learning-based video compression algorithms, MPEG will keep this call open and anticipates future updates to the call.

Interested parties are requested to contact the MPEG AG 5 Convenor Mathias Wien (wien@lfb.rwth-aachen.de) and submit responses for review at the 145th MPEG meeting in January 2024. Further details are given in the call, issued as AG 5 document N 104 and available from the MPEG website.