High Efficiency VVC Video Encoder (CODEO++)
CODEO++ aims to design and implement an efficient video encoder that produces videos stream that meet the specifications of the H.266/Versatile Video Coding (VVC) standard, requiring less computation time (and thus less power consumption), in addition to obtaining a higher compression rate without affecting the subjective quality.
This is based on three fundamental pillars: the use of machine learning techniques to accelerate frame partitioning through fast predictions, acceleration through parallel architectures of the video encoder, and the use of artificial vision techniques to predict the points of interest of the frames and apply perceptual video coding techniques based on these predictions.
Duration: 2022 – 2024 (36 months)
Funding source (Code): Ministerio de Ciencia e Innovación (PID2021-128167OA-I00)
Participant entities: Universidad Politécnica de Madrid
Official webpage: N/A
Laboratory/internal web: N/A
Responsible: Antonio Jesus Diaz Honrubia
Co-Responsible: Guillermo Vigueras González