Variable Bitrate Image Compression with Quality Scaling Factors

Chen, Tong; Ma, Zhan

doi:10.1109/icassp40776.2020.9053885

Cited by 39 publications

(26 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Finally, supporting progressivity or embedded encoding in variable bitrates are well-desired features of future video communication systems. However, while the benefits of using deep learning architectures over traditional methods to achieve progressivity [ 213 ] or supporting variable bitrates [ 214 , 215 ] have been demonstrated for image coding, their application to video is not straightforward due to the inherent complexity of the video coding pipeline.…”

Section: Towards An End-to-end Learning-based Transmission Systemmentioning

confidence: 99%

Machine Learning for Multimedia Communications

Thomos

Maugey

Toni

2022

Sensors

View full text Add to dashboard Cite

Machine learning is revolutionizing the way multimedia information is processed and transmitted to users. After intensive and powerful training, some impressive efficiency/accuracy improvements have been made all over the transmission pipeline. For example, the high model capacity of the learning-based architectures enables us to accurately model the image and video behavior such that tremendous compression gains can be achieved. Similarly, error concealment, streaming strategy or even user perception modeling have widely benefited from the recent learning-oriented developments. However, learning-based algorithms often imply drastic changes to the way data are represented or consumed, meaning that the overall pipeline can be affected even though a subpart of it is optimized. In this paper, we review the recent major advances that have been proposed all across the transmission chain, and we discuss their potential impact and the research challenges that they raise.

show abstract

Section: Towards An End-to-end Learning-based Transmission Systemmentioning

confidence: 99%

Machine Learning for Multimedia Communications

Thomos

Maugey

Toni

2022

Sensors

View full text Add to dashboard Cite

show abstract

“…We use P (•) to denote the probability value on an interval and use P(•) to denote the discretized probability value. The idea of latent scaling [14,15], is to apply a scaling factor s to the latent y. Because the quantization step size is fixed, this leads to a different tradeoff between rate and distortion.…”

Section: Latent Scalingmentioning

confidence: 99%

“…We name this simple single-model variable-bitrate scheme as Naive Scaling. Despite its simplicity, In Section 5 we show that it achieves performance comparable to a more involved scheme [14] that uses K sets of learnable per-channel scaling factors optimized for K Lagrange multipliers β.…”

Section: Nested Quantizationmentioning

confidence: 99%

“…To address this issue, Choi et al [13] proposed a conditional auto-encoder to make the parameters in the encoder and decoder network be dependent on the rate distortion trade-off parameter, β so that a single model can adapt to different rate-distortion tradeoffs. More recently, [14,15,3] proposed latent scaling based variable bitrate compression which is essentially learning to adjust the quantization step size of the latents. After training, the interpolation of the learned scaling parameters can lead to continuous variable bitrate performance.…”

Section: Introductionmentioning

confidence: 99%

“…While variable bitrate solutions make learning based image compression more practical for deployment, they require storing multiple copies of the latent code used for different bitrates, i.e., quality levels [14,15]. In contrast, with progressive coding, the transmitted bitstream is completely embedded, which be truncated at various points and reconstructed into a series of lower bitrate images [18].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Progressive Neural Image Compression With Nested Quantization And Latent Ordering

Zhu

Yang

et al. 2021

2021 IEEE International Conference on Image Processing (ICIP)

View full text Add to dashboard Cite

We present PLONQ, a progressive neural image compression scheme which pushes the boundary of variable bitrate compression by allowing quality scalable coding with a single bitstream. In contrast to existing learned variable bitrate solutions which produce separate bitstreams for each quality, it enables easier rate-control and requires less storage. Leveraging the latent scaling based variable bitrate solution, we introduce nested quantization, a method that defines multiple quantization levels with nested quantization grids, and progressively refines all latents from the coarsest to the finest quantization level. To achieve finer progressiveness in between any two quantization levels, latent elements are incrementally refined with an importance ordering defined in the rate-distortion sense. To the best of our knowledge, PLONQ is the first learningbased progressive image coding scheme and it outperforms SPIHT, a well-known wavelet-based progressive image codec. Index Termsprogressive coding, quality scalable coding, embedded coding, variable bitrate compression, nested quantization

show abstract

An Efficient Image Compression Algorithm Using Neural Networks

Bhattacharjee,

Ghosh

2024

Advances in Geographic Information Science

View full text Add to dashboard Cite

Variable Bitrate Image Compression with Quality Scaling Factors

Cited by 39 publications

References 8 publications

Machine Learning for Multimedia Communications

Machine Learning for Multimedia Communications

Progressive Neural Image Compression With Nested Quantization And Latent Ordering

An Efficient Image Compression Algorithm Using Neural Networks

Contact Info

Product

Resources

About