Patent №
US 11,025,937
Granted
2021-06-01
Filed 2020
Owner
TENCENT AMERICA LLC
Lab
—
AI components
1
vision
Assignment
Recorded
Dataset
AIPD
2023_r1 edition
Application
16819085
A method for video decoding in a decoder is provided. In the method, transform block signaling information is acquired from a coded video bitstream. A transform type is determined based on the transform block signaling information. A low frequency coefficient of one of a plurality of sub transform units is determined based on the transform type and neighboring sub transform units of the one of the plurality of sub transform units. The plurality of sub transform units is partitioned from a current coding block unit (CU). The current coding block unit is subsequently decoded based on low frequency coefficients of the plurality of sub transform units, where the low frequency coefficients include the low frequency coefficient of the one of the plurality of sub transform units.
AI classification
Ownership
TENCENT AMERICA LLC
assignment · 523620493
Assignors
ZHAO, XIN, LI, XIANG, LIU, SHAN
On an employer assignment, the assignors are typically the inventors.