Conditional Residual Coding with Explicit-Implicit Temporal Buffering for Learned Video Compression

Research output: Chapter in book/report/conference proceedingConference contributionResearchpeer review

Authors

Research Organisations

External Research Organisations

  • National Yang Ming Chiao Tung University (NSTC)
View graph of relations

Details

Original languageEnglish
Title of host publication2025 IEEE International Conference on Multimedia and Expo
Subtitle of host publicationJourney to the Center of Machine Imagination, ICME 2025 - Conference Proceedings
PublisherIEEE Computer Society
ISBN (electronic)9798331594954
ISBN (print)979-8-3315-9496-1
Publication statusPublished - 30 Jun 2025
Event2025 IEEE International Conference on Multimedia and Expo, ICME 2025 - Nantes, France
Duration: 30 Jun 20254 Jul 2025

Publication series

NameProceedings - IEEE International Conference on Multimedia and Expo
ISSN (Print)1945-7871
ISSN (electronic)1945-788X

Abstract

This work proposes a hybrid, explicit-implicit temporal buffering scheme for conditional residual video coding. Recent conditional coding methods propagate implicit temporal information for inter-frame coding, demonstrating superior coding performance to those relying exclusively on previously decoded frames (i.e. the explicit temporal information). However, these methods require substantial memory to store a large number of implicit features. This work presents a hybrid buffering strategy. For inter-frame coding, it buffers one previously decoded frame as the explicit temporal reference and a small number of learned features as implicit temporal reference. Our hybrid buffering scheme for conditional residual coding outperforms the single use of explicit or implicit information. Moreover, it allows the total buffer size to be reduced to the equivalent of two video frames with a negligible performance drop on 2K video sequences. The ablation experiment further sheds light on how these two types of temporal references impact the coding performance.

Keywords

    conditional residual coding, implicit and explicit temporal information buffering, Learned video compression

ASJC Scopus subject areas

Cite this

Conditional Residual Coding with Explicit-Implicit Temporal Buffering for Learned Video Compression. / Chen, Yi Hsin; Ho, Kuan Wei; Benjak, Martin et al.
2025 IEEE International Conference on Multimedia and Expo: Journey to the Center of Machine Imagination, ICME 2025 - Conference Proceedings. IEEE Computer Society, 2025. (Proceedings - IEEE International Conference on Multimedia and Expo).

Research output: Chapter in book/report/conference proceedingConference contributionResearchpeer review

Chen, YH, Ho, KW, Benjak, M, Ostermann, J & Peng, WH 2025, Conditional Residual Coding with Explicit-Implicit Temporal Buffering for Learned Video Compression. in 2025 IEEE International Conference on Multimedia and Expo: Journey to the Center of Machine Imagination, ICME 2025 - Conference Proceedings. Proceedings - IEEE International Conference on Multimedia and Expo, IEEE Computer Society, 2025 IEEE International Conference on Multimedia and Expo, ICME 2025, Nantes, France, 30 Jun 2025. https://doi.org/10.1109/ICME59968.2025.11209118, https://doi.org/10.48550/arXiv.2508.01818
Chen, Y. H., Ho, K. W., Benjak, M., Ostermann, J., & Peng, W. H. (2025). Conditional Residual Coding with Explicit-Implicit Temporal Buffering for Learned Video Compression. In 2025 IEEE International Conference on Multimedia and Expo: Journey to the Center of Machine Imagination, ICME 2025 - Conference Proceedings (Proceedings - IEEE International Conference on Multimedia and Expo). IEEE Computer Society. https://doi.org/10.1109/ICME59968.2025.11209118, https://doi.org/10.48550/arXiv.2508.01818
Chen YH, Ho KW, Benjak M, Ostermann J, Peng WH. Conditional Residual Coding with Explicit-Implicit Temporal Buffering for Learned Video Compression. In 2025 IEEE International Conference on Multimedia and Expo: Journey to the Center of Machine Imagination, ICME 2025 - Conference Proceedings. IEEE Computer Society. 2025. (Proceedings - IEEE International Conference on Multimedia and Expo). doi: 10.1109/ICME59968.2025.11209118, 10.48550/arXiv.2508.01818
Chen, Yi Hsin ; Ho, Kuan Wei ; Benjak, Martin et al. / Conditional Residual Coding with Explicit-Implicit Temporal Buffering for Learned Video Compression. 2025 IEEE International Conference on Multimedia and Expo: Journey to the Center of Machine Imagination, ICME 2025 - Conference Proceedings. IEEE Computer Society, 2025. (Proceedings - IEEE International Conference on Multimedia and Expo).
Download
@inproceedings{1c645034d42d4c89a69f4a124e7e758e,
title = "Conditional Residual Coding with Explicit-Implicit Temporal Buffering for Learned Video Compression",
abstract = "This work proposes a hybrid, explicit-implicit temporal buffering scheme for conditional residual video coding. Recent conditional coding methods propagate implicit temporal information for inter-frame coding, demonstrating superior coding performance to those relying exclusively on previously decoded frames (i.e. the explicit temporal information). However, these methods require substantial memory to store a large number of implicit features. This work presents a hybrid buffering strategy. For inter-frame coding, it buffers one previously decoded frame as the explicit temporal reference and a small number of learned features as implicit temporal reference. Our hybrid buffering scheme for conditional residual coding outperforms the single use of explicit or implicit information. Moreover, it allows the total buffer size to be reduced to the equivalent of two video frames with a negligible performance drop on 2K video sequences. The ablation experiment further sheds light on how these two types of temporal references impact the coding performance.",
keywords = "conditional residual coding, implicit and explicit temporal information buffering, Learned video compression",
author = "Chen, {Yi Hsin} and Ho, {Kuan Wei} and Martin Benjak and J{\"o}rn Ostermann and Peng, {Wen Hsiao}",
note = "Publisher Copyright: {\textcopyright} 2025 IEEE.; 2025 IEEE International Conference on Multimedia and Expo, ICME 2025, ICME 2025 ; Conference date: 30-06-2025 Through 04-07-2025",
year = "2025",
month = jun,
day = "30",
doi = "10.1109/ICME59968.2025.11209118",
language = "English",
isbn = "979-8-3315-9496-1",
series = "Proceedings - IEEE International Conference on Multimedia and Expo",
publisher = "IEEE Computer Society",
booktitle = "2025 IEEE International Conference on Multimedia and Expo",
address = "United States",

}

Download

TY - GEN

T1 - Conditional Residual Coding with Explicit-Implicit Temporal Buffering for Learned Video Compression

AU - Chen, Yi Hsin

AU - Ho, Kuan Wei

AU - Benjak, Martin

AU - Ostermann, Jörn

AU - Peng, Wen Hsiao

N1 - Publisher Copyright: © 2025 IEEE.

PY - 2025/6/30

Y1 - 2025/6/30

N2 - This work proposes a hybrid, explicit-implicit temporal buffering scheme for conditional residual video coding. Recent conditional coding methods propagate implicit temporal information for inter-frame coding, demonstrating superior coding performance to those relying exclusively on previously decoded frames (i.e. the explicit temporal information). However, these methods require substantial memory to store a large number of implicit features. This work presents a hybrid buffering strategy. For inter-frame coding, it buffers one previously decoded frame as the explicit temporal reference and a small number of learned features as implicit temporal reference. Our hybrid buffering scheme for conditional residual coding outperforms the single use of explicit or implicit information. Moreover, it allows the total buffer size to be reduced to the equivalent of two video frames with a negligible performance drop on 2K video sequences. The ablation experiment further sheds light on how these two types of temporal references impact the coding performance.

AB - This work proposes a hybrid, explicit-implicit temporal buffering scheme for conditional residual video coding. Recent conditional coding methods propagate implicit temporal information for inter-frame coding, demonstrating superior coding performance to those relying exclusively on previously decoded frames (i.e. the explicit temporal information). However, these methods require substantial memory to store a large number of implicit features. This work presents a hybrid buffering strategy. For inter-frame coding, it buffers one previously decoded frame as the explicit temporal reference and a small number of learned features as implicit temporal reference. Our hybrid buffering scheme for conditional residual coding outperforms the single use of explicit or implicit information. Moreover, it allows the total buffer size to be reduced to the equivalent of two video frames with a negligible performance drop on 2K video sequences. The ablation experiment further sheds light on how these two types of temporal references impact the coding performance.

KW - conditional residual coding

KW - implicit and explicit temporal information buffering

KW - Learned video compression

UR - http://www.scopus.com/inward/record.url?scp=105022664047&partnerID=8YFLogxK

U2 - 10.1109/ICME59968.2025.11209118

DO - 10.1109/ICME59968.2025.11209118

M3 - Conference contribution

AN - SCOPUS:105022664047

SN - 979-8-3315-9496-1

T3 - Proceedings - IEEE International Conference on Multimedia and Expo

BT - 2025 IEEE International Conference on Multimedia and Expo

PB - IEEE Computer Society

T2 - 2025 IEEE International Conference on Multimedia and Expo, ICME 2025

Y2 - 30 June 2025 through 4 July 2025

ER -

By the same author(s)