Image Compression Based on Hybrid Domain Attention and Postprocessing Enhancement

Deep learning-based image compression methods have made significant achievements recently, of which the two key components are the entropy model for latent representations and the encoder-decoder network. Both the inaccurate estimation of the entropy estimation model and the existence of information...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Computational intelligence and neuroscience Ročník 2022; s. 1 - 12
Hlavní autoři: Bao, Yuting, Tao, Yuwen, Qian, Pengjiang
Médium: Journal Article
Jazyk:angličtina
Vydáno: United States Hindawi 17.03.2022
John Wiley & Sons, Inc
Témata:
Algorithms
Analysis
Coders
Compression
Computational linguistics
Computer peripherals industry
Data Compression
Decoding
Deep learning
Design
Domains
Efficiency
Electric Power Supplies
Encoders-Decoders
Entropy
Human resource management
Image coding
Image compression
Image enhancement
Information technology
Language processing
Measurement
Modules
Natural language interfaces
Neural networks
Neural Networks, Computer
Normal Distribution
Probabilistic models
Propagation
Redundancy
Representations
Wavelet transforms
United States
ISSN:1687-5265, 1687-5273, 1687-5273
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Deep learning-based image compression methods have made significant achievements recently, of which the two key components are the entropy model for latent representations and the encoder-decoder network. Both the inaccurate estimation of the entropy estimation model and the existence of information redundancy in latent representations lead to a reduction in the compression efficiency. To address these issues, the study suggests an image compression method based on a hybrid domain attention mechanism and postprocessing improvement. This study embeds hybrid domain attention modules as nonlinear transformers in both the main encoder-decoder network and the hyperprior network, aiming at constructing more compact latent features and hyperpriors and then model the latent features as parametric Gaussian-scale mixture models to obtain more precise entropy estimation. In addition, we propose a solution to the errors introduced by quantization in image compression by adding an inverse quantization module. On the decoding side, we also provide a postprocessing enhancement module to further increase image compression performance. The experimental results show that the peak signal-to-noise rate (PSNR) and multiscale structural similarity (MS-SSIM) of the proposed method are higher than those of traditional compression methods and advanced neural network-based methods.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
content type line 23
Academic Editor: Xin Ning
ISSN:1687-5265
1687-5273
1687-5273
DOI:10.1155/2022/4926124