Skip to main navigation Skip to search Skip to main content

CoordFlow: Coordinate Flow for Pixel-Wise Neural Video Representation

Daniel Silver, Ron Kimmel

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

CoordFlow is a pixel-wise Implicit Neural Representation (INR) framework for video compression, achieving state-of-the-art results among pixel-wise methods and rivaling both frame-wise and classic techniques. By segmenting videos into layers, each represented by specialized neural networks with built-in motion compensation, CoordFlow boosts performance and enables additional practical video tasks.

Original languageEnglish
Title of host publicationProceedings - DCC 2025
Subtitle of host publication2025 Data Compression Conference
EditorsAli Bilgin, James E. Fowler, Joan Serra-Sagrista, Yan Ye, James A. Storer
Pages397
Number of pages1
ISBN (Electronic)9798331534714
DOIs
StatePublished - 2025
Externally publishedYes
Event2025 Data Compression Conference, DCC 2025 - Snowbird, United States
Duration: 18 Mar 202521 Mar 2025

Publication series

NameData Compression Conference Proceedings

Conference

Conference2025 Data Compression Conference, DCC 2025
Country/TerritoryUnited States
CitySnowbird
Period18/03/2521/03/25

Keywords

  • implicit neural representation
  • video compression

All Science Journal Classification (ASJC) codes

  • Computer Networks and Communications

Fingerprint

Dive into the research topics of 'CoordFlow: Coordinate Flow for Pixel-Wise Neural Video Representation'. Together they form a unique fingerprint.

Cite this