Signaling for illumination compensation

Abstract

A video coder that implements illumination compensation is provided. The video coder receives a first block of pixels in a first video picture to be coded as a current block, wherein the current block is associated with a motion vector that references a second block of pixels in a second video picture as a reference block. The video coder performs inter-prediction for the current block by using the motion vector to generate a set of motion-compensated pixels for the current block. The video coder modifies the set of motion-compensated pixels of the current block by applying a linear model that is computed based on neighboring samples of the reference block and of the current block. The neighboring samples are identified based on a position of the current block within a larger block.

1 . A method comprising:
 determining if size of a current CU is larger than a threshold;   in the case that the size of the current CU is larger than the threshold, the current CU is split into a plurality of sub-CUs;   encoding or decoding the sub-CUs using Bi-directional optical flow.   
     
     
         2 . The method of  claim 1 , wherein the current CU is split into the sub-CUs with a pre-determined size. 
     
     
         3 . The method of  claim 1 , wherein the pre-determined size is equal to the threshold. 
     
     
         4 . An apparatus for video coding, the apparatus comprising one or more electronic circuits or processors arranged to:
 determine if size of a current CU is larger than a threshold;   in the case that the size of the current CU is larger than the threshold, the current CU is split into a plurality of sub-CUs;   encoding or decoding the sub-CUs using Bi-directional optical flow.