BLOCK-BASED MOTION ESTIMATION FOR DEEP-LEARNED VIDEO CODING

被引:1
作者
Pientka, Sophie [1 ]
Schaefer, Michael [1 ]
Pfaff, Jonathan [1 ]
Schwarz, Heiko [1 ,2 ]
Marpe, Detlev [1 ]
Wiegand, Thomas [1 ,3 ]
机构
[1] Heinrich Hertz Inst Nachrichtentech Berlin GmbH, Fraunhofer Inst Telecommun, Berlin, Germany
[2] Free Univ Berlin, Inst Comp Sci, Berlin, Germany
[3] Tech Univ Berlin, Dept Telecommun Syst, Berlin, Germany
来源
2023 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, ICIP | 2023年
关键词
video compression; variational autoencoders; block matching; motion estimation; motion compensation; ALGORITHM;
D O I
10.1109/ICIP49359.2023.10222411
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
The research on deep-learned end-to-end video compression has attracted a lot of attention over the course of recent years. A central component of many approaches is to perform motion-compensated prediction by using convolutional neural networks (CNN) which determine a compressed representation of the motion field as features. Often, this task is divided into searching motion vectors by one network and efficiently representing them by another one. However, these networks may find motion fields far from optimal because the search radius of CNNs is mainly determined by their depth and kernel size. In this paper, we apply motion estimation techniques from classical block-based hybrid video compression to search a motion field which is then fed into a variational autoencoder. These strategies include different distortion measures, different block partitions and an improved approximation of the residual bitrate. With our modifications, bitrate savings of up to 13% over the underlying end-to-end based video codec can be obtained.
引用
收藏
页码:3444 / 3448
页数:5
相关论文
共 25 条
[1]   Scale-space flow for end-to-end optimized video compression [J].
Agustsson, Eirikur ;
Minnen, David ;
Johnston, Nick ;
Balle, Johannes ;
Hwang, Sung Jin ;
Toderici, George .
2020 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2020), 2020, :8500-8509
[2]  
[Anonymous], 2020, H266 ITUT
[3]  
[Anonymous], 2013, H265 ITUT
[4]  
[Anonymous], 2003, ITU-T Rec. H.264,
[5]   Smoothness Constraints in Recursive Search Motion Estimation for Picture Rate Conversion [J].
Bartels, Chris ;
de Haan, Gerard .
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, 2010, 20 (10) :1310-1319
[6]  
Bjontegaard G., 2001, VCEGM33
[7]  
Boyce J., 2018, JVETJ1010
[8]  
Browne A., 2021, JVETW2002
[9]   Motion Vector Coding and Block Merging in the Versatile Video Coding Standard [J].
Chien, Wei-Jung ;
Zhang, Li ;
Winken, Martin ;
Li, Xiang ;
Liao, Ru-Ling ;
Gao, Han ;
Hsu, Chih-Wei ;
Liu, Hongbin ;
Chen, Chun-Chi .
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, 2021, 31 (10) :3848-3861
[10]   Novel True-Motion Estimation Algorithm and Its Application to Motion-Compensated Temporal Frame Interpolation [J].
Dikbas, Salih ;
Altunbasak, Yucel .
IEEE TRANSACTIONS ON IMAGE PROCESSING, 2013, 22 (08) :2931-2945