Methods and apparatuses for prediction refinement with optical flow

Abstract

A bit-depth representation method of prediction refinement with optical flow (PROF), apparatus, and a non-transitory computer-readable storage medium are provided. The method includes obtaining a first reference picture associated with a video block within the video signal and a first motion vector (MV) from the video block in a current picture to a reference block in the first reference picture, obtaining the first prediction samples I(i,j)'s of the video block, controlling internal bit-depths of the internal PROF parameters, obtaining prediction refinement values for the first prediction samples I(i,j)'s based on horizontal and vertical gradient values and horizontal and vertical motion differences, obtaining second prediction samples I'(i,j)'s associated with the second MV and corresponding prediction refinement values for the second prediction samples I′(i,j)'s, and obtaining final prediction samples based on the combination of the first prediction samples I(i,j)'s, second prediction samples I′(i,j)'s, and the prediction refinement values.