Back to Search
Document details
CE2: Subsampled sum-modified-Laplacian (Test 4.1.1)
Abstract
This contribution reports the experimental results of CE2-4.1.1 test proposed in call for proposal (CfP) response document of JVET-J0013. In CE2-4.1.1 test, 4x4 block classification based on subsampled sum-modified-Laplacian (SSML) was tested in order to reduce the computational complexity of the block classification of adaptive loop filter (ALF) in benchmark set (BMS), especially for worst case operation count. The CE2-4.1.1 method reportedly reduces approximately 78% of worst case operation count compared to block classification compared to ALF in BMS. The experimental results reportedly show average BD-rates of 0.19%, 0.18%, and 0.15% with decoding time of 96%, 96%, and 96% for AI, RA, and LDB cases, respectively, compared to VVC test model (VTM) 1.0 anchor with ALF. And, the experimental results reportedly show average BD-rates of 0.15%, 0.17%, and 0.10% with decoding time of 97%, 97%, and 97% for AI, RA, and LDB cases compared to BMS 1.0 anchor.
JVET-K0164 CE2: Subsampled sum-modified-Laplacian (Test 4.1.1) [S.-C. Lim, J. Kang, H. Lee, J. Lee, S. Cho, H. Y. Kim (ETRI)]
Note: This document was withdrawn by mistake, and the problem was later corrected.
References:
PATENTS:
CN109716345B
0.34
2023-09-15
CN115018728A
0.32
2022-09-06
CN106901741A
0.30
2017-06-30
CN118643524A
0.28
2024-09-13
WO2014187432A1
0.26
2014-11-27
WO1994011987A1
0.24
1994-05-26
US20250075202A1
0.22
2025-03-06
CN120912562A
0.20
2025-11-07
JVET-K0164.docx
JVET-K0164_r1.docx
JVET-K0164_results_CE2.4.1.1 + BMS1.0 vs. BMS1.0 (2x2BC+SSML).xlsm
JVET-K0164_results_CE2.4.1.1 + BMS1.0 vs. BMS1.0_r1 (4x4BC+SSML).xlsm
JVET-K0164_results_CE2.4.1.1 + VTM1.0 vs. VTM1.0 (2x2BC+SSML).xlsm
JVET-K0164_results_CE2.4.1.1 + VTM1.0 vs. VTM1.0_r1 (4x4BC+SSML).xlsm
Decisions
withdrawn
Note: This document was withdrawn by mistake, and the problem was later corrected.
Citation