US 9,813,708 B2
Method and system of coding prediction for screen video
Chun-Lung Lin, Yunlin County (TW); and Pei-Hsuan Tsai, Taipei (TW)
Assigned to INDUSTRIAL TECHNOLOGY RESEARCH INSTITUTE, Hsinchu (TW)
Filed by INDUSTRIAL TECHNOLOGY RESEARCH INSTITUTE, Hsinchu (TW)
Filed on Apr. 17, 2014, as Appl. No. 14/255,261.
Claims priority of application No. 102144619 A (TW), filed on Dec. 5, 2013.
Prior Publication US 2015/0163485 A1, Jun. 11, 2015
Int. Cl. H04N 19/583 (2014.01); H04N 19/105 (2014.01); H04N 19/56 (2014.01); H04N 19/52 (2014.01); H04N 19/176 (2014.01); H04N 19/117 (2014.01); H04N 19/14 (2014.01); H04N 19/567 (2014.01)
CPC H04N 19/105 (2014.11) [H04N 19/117 (2014.11); H04N 19/14 (2014.11); H04N 19/176 (2014.11); H04N 19/52 (2014.11); H04N 19/56 (2014.11); H04N 19/567 (2014.11)] 17 Claims
OG exemplary drawing
 
1. A method of coding prediction for screen video, comprising:
using a classifier to classify a plurality of coding blocks of a screen video contents into a plurality of block types; and
based on the plurality of block types corresponding to the plurality of coding blocks, using a computing device to filter at least one of a plurality of candidate blocks from the plurality of coding blocks, to compute a first candidate motion vector set of a type-based motion merge (TMM) mode, and to compute a second candidate motion vector set of a type-based advanced motion vector prediction (AMVP) mode simultaneously with computing the first candidate motion vector set,
wherein each of the at least one of the plurality of candidate blocks has a block type different from that of a current coding block,
wherein computing the first candidate motion vector set of the TMM mode including:
sorting the plurality of candidate blocks by using the plurality of block types comprising a text image and a natural image;
constructing a candidate set of TMM mode; and
using the plurality of block types corresponding to the plurality of candidate blocks in the candidate set to filter out at least one candidate motion vector corresponding to the at least one of the plurality of candidate blocks, then selecting one or more best matched motion vectors to form the first candidate motion vector set.