Back to Search Document details
15th Meeting: Gothenburg, SE, July 2019 2019-06-17 15:42
CE5-3.1 Sub-sample MV threshold for deblocking decisions
Authors: K. Andersson Ericsson J. Enhorn Ericsson
Abstract
This contribution reports results of CE5-3.1. VTM-5.0 fails to remove blocking artifacts that origins from differences smaller than one sample for at least one motion vector component on respective side of a prediction block boundary since the threshold for deblocking decision is one sample. In this core experiment a threshold of half-sample can combat most visual artifacts occurring from small differences in motion in neighbouring blocks.
Meeting Notes
JVET-O0061 CE5-3.1 Sub-sample MV threshold for deblocking decisions [K. Andersson, J. Enhorn (Ericsson)]
Cited:
JVET-P0002 JVET-P0003 JVET-P0071 JVET-P0074 JVET-P0075 JVET-P0113 JVET-P0114 JVET-P0115 JVET-P0116 JVET-P0117 JVET-P0118 JVET-P0126 JVET-P0150 JVET-P0151 JVET-P0155 JVET-P0162 JVET-P0168 JVET-P0220 JVET-P0234 JVET-P0236 more than 20
ARCHIVE :
CE5_3.1.XLSM CE5_3.1_No_ALF.XLSM JVET-O0061-v1.docx
Decisions
JVET-O0061 CE5-3.1 Sub-sample MV threshold for deblocking decisions [K. Andersson, J. Enhorn (Ericsson)]
Citation