Information Bottleneck Driven Deep Video Compression—IBOpenDVCW

Video compression remains a challenging task despite significant advancements in end-to-end optimized deep networks for video coding. This study, inspired by information bottleneck (IB) theory, introduces a novel approach that combines IB theory with wavelet transform. We perform a comprehensive ana...

Full description

Saved in:
Bibliographic Details
Main Authors: Timor Leiderman, Yosef Ben Ezra
Format: Article
Language:English
Published: MDPI AG 2024-09-01
Series:Entropy
Subjects:
Online Access:https://www.mdpi.com/1099-4300/26/10/836
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850205789071867904
author Timor Leiderman
Yosef Ben Ezra
author_facet Timor Leiderman
Yosef Ben Ezra
author_sort Timor Leiderman
collection DOAJ
description Video compression remains a challenging task despite significant advancements in end-to-end optimized deep networks for video coding. This study, inspired by information bottleneck (IB) theory, introduces a novel approach that combines IB theory with wavelet transform. We perform a comprehensive analysis of information and mutual information across various mother wavelets and decomposition levels. Additionally, we replace the conventional average pooling layers with a discrete wavelet transform creating more advanced pooling methods to investigate their effects on information and mutual information. Our results demonstrate that the proposed model and training technique outperform existing state-of-the-art video compression methods, delivering competitive rate-distortion performance compared to the AVC/H.264 and HEVC/H.265 codecs.
format Article
id doaj-art-478b878a1ae04f02b490a36bc80fb0b8
institution OA Journals
issn 1099-4300
language English
publishDate 2024-09-01
publisher MDPI AG
record_format Article
series Entropy
spelling doaj-art-478b878a1ae04f02b490a36bc80fb0b82025-08-20T02:11:00ZengMDPI AGEntropy1099-43002024-09-01261083610.3390/e26100836Information Bottleneck Driven Deep Video Compression—IBOpenDVCWTimor Leiderman0Yosef Ben Ezra1Faculty of Electrical Engineering, Holon Institute of Technology, 52 Golomb Str., P.O. Box 305, Holon 58102, IsraelFaculty of Electrical Engineering, Holon Institute of Technology, 52 Golomb Str., P.O. Box 305, Holon 58102, IsraelVideo compression remains a challenging task despite significant advancements in end-to-end optimized deep networks for video coding. This study, inspired by information bottleneck (IB) theory, introduces a novel approach that combines IB theory with wavelet transform. We perform a comprehensive analysis of information and mutual information across various mother wavelets and decomposition levels. Additionally, we replace the conventional average pooling layers with a discrete wavelet transform creating more advanced pooling methods to investigate their effects on information and mutual information. Our results demonstrate that the proposed model and training technique outperform existing state-of-the-art video compression methods, delivering competitive rate-distortion performance compared to the AVC/H.264 and HEVC/H.265 codecs.https://www.mdpi.com/1099-4300/26/10/836deep video compressionwaveletsinformation bottleneckneural networks
spellingShingle Timor Leiderman
Yosef Ben Ezra
Information Bottleneck Driven Deep Video Compression—IBOpenDVCW
Entropy
deep video compression
wavelets
information bottleneck
neural networks
title Information Bottleneck Driven Deep Video Compression—IBOpenDVCW
title_full Information Bottleneck Driven Deep Video Compression—IBOpenDVCW
title_fullStr Information Bottleneck Driven Deep Video Compression—IBOpenDVCW
title_full_unstemmed Information Bottleneck Driven Deep Video Compression—IBOpenDVCW
title_short Information Bottleneck Driven Deep Video Compression—IBOpenDVCW
title_sort information bottleneck driven deep video compression ibopendvcw
topic deep video compression
wavelets
information bottleneck
neural networks
url https://www.mdpi.com/1099-4300/26/10/836
work_keys_str_mv AT timorleiderman informationbottleneckdrivendeepvideocompressionibopendvcw
AT yosefbenezra informationbottleneckdrivendeepvideocompressionibopendvcw