Back to Search Document details
32nd Meeting: Hannover, DE, October 2023 2023-10-12 11:19
EE1-2.1: Deep Reference Frame Generation for Inter Prediction Enhancement
Authors: W. Bao Wuhan Univ. Xinxin Chen Wuhan Univ. J. Jia Wuhan Univ. Yuhuai Zhang Wuhan Univ. Zongxie Chen Wuhan Univ. Z. Liu Tencent Xiaozhong Xu Tencent Shan Liu Tencent Zhenzhong Chen
Abstract
This contribution reports the EE1-2.1 results of NN-based inter prediction. Based on JVET-AE0112 [1], this contribution has simplified the network architecture to reduce computational complexity and has accelerated float32 SADL inference. this method can achieve bitrate savings for the Y/U/V components under the NNVC-RA configuration
Meeting Notes
JVET-AF0208 EE1-2.1: Deep Reference Frame Generation for Inter Prediction Enhancement W. Bao, X. Chen, J. Jia, Y. Zhang, Z. Chen (Wuhan Univ.), Z. Liu, X. Xu, S. Liu (Tencent) [late]

See JVET-AF0023.

Cited:
JVET-AF0268 X
ARCHIVE :
JVET-AF0208-v3.docx NNVC-6.0(wo_intra_lop)_vs_EE1-2.1.xlsm NNVC-6.0_vs_EE1-2.1.xlsm ~$ET-AF0208-v2.docx
Decisions
See JVET-AF0023.
Citation