US 12,149,708 B2
Machine learning of encoding parameters for a network using a video encoder
Ravi Kumar Boddeti, Hyderabad (IN); Vinayak Pore, Pune (IN); Hassane Samir Azar, Los Altos, CA (US); and Prashant Sohani, Pune (IN)
Assigned to NVIDIA Corporation, Santa Clara, CA (US)
Filed by NVIDIA Corporation, Santa Clara, CA (US)
Filed on Aug. 16, 2021, as Appl. No. 17/402,953.
Prior Publication US 2023/0048189 A1, Feb. 16, 2023
Int. Cl. H04N 19/00 (2014.01); G06N 20/00 (2019.01); H04N 19/127 (2014.01); H04N 19/142 (2014.01); H04N 19/166 (2014.01); H04N 19/65 (2014.01)
CPC H04N 19/166 (2014.11) [G06N 20/00 (2019.01); H04N 19/127 (2014.11); H04N 19/142 (2014.11); H04N 19/65 (2014.11)] 20 Claims
OG exemplary drawing
 
1. A method comprising:
transmitting data to cause an encoder to use at least one first value of at least one encoding parameter to convert first video data into first encoded video data;
receiving, using a network client, feedback associated with a receipt of a stream comprising the first encoded video data, the feedback corresponding to network performance associated with an encoded bitrate of the first encoded video data in the stream;
predicting, using a Machine Learning Model (MLM) and the feedback, at least one second value of the at least one encoding parameter representing a prediction of a target bitrate that penalizes a mismatch between the target bitrate and an actual bitrate of second encoded video data and accounts for one or more criteria corresponding to the network performance;
transmitting data to cause the encoder to use the at least one second value of the at least one encoding parameter predicted using the MLM to encode video data of the stream into the second encoded video data using the target bitrate; and
streaming the second encoded video data over a network channel associated with the network client.