US 11,997,283 B2
Machine learning based content-aware image frame encoding
Sankar Shanmugam, Bangalore (IN); and Ashwin Kulkarni, Bangalore (IN)
Assigned to SONY GROUP CORPORATION, Tokyo (JP)
Filed by SONY GROUP CORPORATION, Tokyo (JP)
Filed on Mar. 30, 2022, as Appl. No. 17/708,369.
Prior Publication US 2023/0319285 A1, Oct. 5, 2023
Int. Cl. H04N 19/136 (2014.01); G06T 5/70 (2024.01); G06T 7/50 (2017.01); G06V 20/52 (2022.01); H04N 19/124 (2014.01); H04N 19/154 (2014.01); H04N 19/162 (2014.01); H04N 19/176 (2014.01); H04N 19/42 (2014.01)
CPC H04N 19/136 (2014.11) [G06T 5/70 (2024.01); G06T 7/50 (2017.01); G06V 20/52 (2022.01); H04N 19/124 (2014.11); H04N 19/154 (2014.11); H04N 19/162 (2014.11); H04N 19/176 (2014.11); H04N 19/42 (2014.11); G06T 2207/20081 (2013.01); G06T 2207/30232 (2013.01)] 20 Claims
OG exemplary drawing
 
1. An electronic device, comprising:
circuitry configured to:
detect at least one foreground region and at least one background region from a first image frame;
determine, from the first image frame, a set of first macroblocks associated with the detected at least one foreground region and a set of second macroblocks associated with the detected at least one background region;
determine a bit allocation control parameter associated with the determined set of second macroblocks, wherein the bit allocation control parameter comprises a quantization parameter;
update the determined bit allocation control parameter associated with the determined set of second macroblocks based on an application of a first trained machine learning model;
determine, as the updated bit allocation control parameter, an updated quantization parameter between a minimum threshold quantization parameter and a default quantization parameter based on the application of the first trained machine learning model; and
encode the first image frame based on the updated bit allocation control parameter to obtain a second image frame,
wherein a first image quality index associated with the first image frame matches a second image quality index associated with the second image frame within a threshold range.