Rate-Distortion Optimized Encoding for Deep Image Compression

Deep-learned variational auto-encoders (VAE) have shown remarkable capabilities for lossy image compression. These neural networks typically employ non-linear convolutional layers for finding a compressible representation of the input image. Advanced techniques such as vector quantization, context-a...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Michael Schafer, Sophie Pientka, Jonathan Pfaff, Heiko Schwarz, Detlev Marpe, Thomas Wiegand
Formato: article
Lenguaje:EN
Publicado: IEEE 2021
Materias:
Acceso en línea:https://doaj.org/article/bb87eff80345459faf4e5ddac9532e1c
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:bb87eff80345459faf4e5ddac9532e1c
record_format dspace
spelling oai:doaj.org-article:bb87eff80345459faf4e5ddac9532e1c2021-11-23T00:02:21ZRate-Distortion Optimized Encoding for Deep Image Compression2644-122510.1109/OJCAS.2021.3124995https://doaj.org/article/bb87eff80345459faf4e5ddac9532e1c2021-01-01T00:00:00Zhttps://ieeexplore.ieee.org/document/9623337/https://doaj.org/toc/2644-1225Deep-learned variational auto-encoders (VAE) have shown remarkable capabilities for lossy image compression. These neural networks typically employ non-linear convolutional layers for finding a compressible representation of the input image. Advanced techniques such as vector quantization, context-adaptive arithmetic coding and variable-rate compression have been implemented in these auto-encoders. Notably, these networks rely on an end-to-end approach, which fundamentally differs from hybrid, block-based video coding systems. Therefore, signal-dependent encoder optimizations have not been thoroughly investigated for VAEs yet. However, rate-distortion optimized encoding heavily determines the compression performance of state-of-the-art video codecs. Designing such optimizations for non-linear, multi-layered networks requires to understand the relationship between the quantization, the bit allocation of the features and the distortion. Therefore, this paper examines the rate-distortion performance of a variable-rate VAE. In particular, one demonstrates that the trained encoder network typically finds features with a near-optimal bit allocation across the channels. Furthermore, one approximates the relationship between distortion and quantization by a higher-order polynomial, whose coefficients can be robustly estimated. Based on these considerations, the authors investigate an encoding algorithm for the Lagrange optimization, which significantly improves the coding efficiency.Michael SchaferSophie PientkaJonathan PfaffHeiko SchwarzDetlev MarpeThomas WiegandIEEEarticleDeep image compressionvariational auto-encodersrate-distortion optimized encodingnon-linear transform codingElectric apparatus and materials. Electric circuits. Electric networksTK452-454.4ENIEEE Open Journal of Circuits and Systems, Vol 2, Pp 633-647 (2021)
institution DOAJ
collection DOAJ
language EN
topic Deep image compression
variational auto-encoders
rate-distortion optimized encoding
non-linear transform coding
Electric apparatus and materials. Electric circuits. Electric networks
TK452-454.4
spellingShingle Deep image compression
variational auto-encoders
rate-distortion optimized encoding
non-linear transform coding
Electric apparatus and materials. Electric circuits. Electric networks
TK452-454.4
Michael Schafer
Sophie Pientka
Jonathan Pfaff
Heiko Schwarz
Detlev Marpe
Thomas Wiegand
Rate-Distortion Optimized Encoding for Deep Image Compression
description Deep-learned variational auto-encoders (VAE) have shown remarkable capabilities for lossy image compression. These neural networks typically employ non-linear convolutional layers for finding a compressible representation of the input image. Advanced techniques such as vector quantization, context-adaptive arithmetic coding and variable-rate compression have been implemented in these auto-encoders. Notably, these networks rely on an end-to-end approach, which fundamentally differs from hybrid, block-based video coding systems. Therefore, signal-dependent encoder optimizations have not been thoroughly investigated for VAEs yet. However, rate-distortion optimized encoding heavily determines the compression performance of state-of-the-art video codecs. Designing such optimizations for non-linear, multi-layered networks requires to understand the relationship between the quantization, the bit allocation of the features and the distortion. Therefore, this paper examines the rate-distortion performance of a variable-rate VAE. In particular, one demonstrates that the trained encoder network typically finds features with a near-optimal bit allocation across the channels. Furthermore, one approximates the relationship between distortion and quantization by a higher-order polynomial, whose coefficients can be robustly estimated. Based on these considerations, the authors investigate an encoding algorithm for the Lagrange optimization, which significantly improves the coding efficiency.
format article
author Michael Schafer
Sophie Pientka
Jonathan Pfaff
Heiko Schwarz
Detlev Marpe
Thomas Wiegand
author_facet Michael Schafer
Sophie Pientka
Jonathan Pfaff
Heiko Schwarz
Detlev Marpe
Thomas Wiegand
author_sort Michael Schafer
title Rate-Distortion Optimized Encoding for Deep Image Compression
title_short Rate-Distortion Optimized Encoding for Deep Image Compression
title_full Rate-Distortion Optimized Encoding for Deep Image Compression
title_fullStr Rate-Distortion Optimized Encoding for Deep Image Compression
title_full_unstemmed Rate-Distortion Optimized Encoding for Deep Image Compression
title_sort rate-distortion optimized encoding for deep image compression
publisher IEEE
publishDate 2021
url https://doaj.org/article/bb87eff80345459faf4e5ddac9532e1c
work_keys_str_mv AT michaelschafer ratedistortionoptimizedencodingfordeepimagecompression
AT sophiepientka ratedistortionoptimizedencodingfordeepimagecompression
AT jonathanpfaff ratedistortionoptimizedencodingfordeepimagecompression
AT heikoschwarz ratedistortionoptimizedencodingfordeepimagecompression
AT detlevmarpe ratedistortionoptimizedencodingfordeepimagecompression
AT thomaswiegand ratedistortionoptimizedencodingfordeepimagecompression
_version_ 1718417398058450944