US 12,382,069 B2
Microdosing for low bitrate video compression
Abdelaziz Djelouah, Zürich (CH); Leonhard Markus Helminger, Zurich (CH); Roberto Gerson De Albuquerque Azevedo, Zurich (CH); Christopher Richard Schroers, Uster (CH); Scott Labrozzi, Cary, NC (US); and Yuanyi Xue, Alameda, CA (US)
Assigned to Disney Enterprises, Inc., Burbank, CA (US); and ETH ZÜRICH (EIDGENÖSSISCHE TECHNISCHE HOCHSCHULE ZÜRICH), Zurich (CH)
Filed by Disney Enterprises, Inc., Burbank, CA (US); and ETH Zürich (EIDGENÖSSISCHE TECHNISCHE HOCHSCHULE ZÜRICH), Zürich (CH)
Filed on May 2, 2024, as Appl. No. 18/653,776.
Application 18/653,776 is a continuation of application No. 17/704,722, filed on Mar. 25, 2022, granted, now 12,010,335.
Claims priority of provisional application 63/255,280, filed on Oct. 13, 2021.
Claims priority of provisional application 63/172,315, filed on Apr. 8, 2021.
Prior Publication US 2024/0283957 A1, Aug. 22, 2024
This patent is subject to a terminal disclaimer.
Int. Cl. H04N 19/42 (2014.01); H04N 19/124 (2014.01); H04N 19/176 (2014.01)
CPC H04N 19/42 (2014.11) 17 Claims
OG exemplary drawing
 
1. A method for use by a machine learning (ML) model-based video decoder, the method comprising:
receiving, by a degradation-aware block based Micro-Residual-Network (MicroRN) defined by a number of hidden channels and a number of degradation-aware blocks of the MicroRN, a first compressed video frame subset;
receiving, by the MicroRN, first decompression data for the first compressed video frame subset;
decoding, by the MicroRN, the first compressed video frame subset using the first decompression data;
receiving, by the MicroRN, a second compressed video frame subset;
receiving, by the MicroRN, second decompression data for the second compressed video frame subset; and
decoding, by the MicroRN, the second compressed video frame subset using the second decompression data, without utilizing a residual network of a generative adversarial network (GAN) trained decoder.