Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

SLIC: A Learned Image Codec Using Structure and Color

Srivatsa Prativadibhayankaram, Mahadev Prasad Panda, Thomas Richter, Heiko Sparenberg, Siegfried Fößel, André Kaup

TL;DR

The structure and color based learned image codec (SLIC) is proposed in which the task of compression is split into that of luminance and chrominance and the advantages of the method are illustrated through the visualization of channel impulse responses, latent channels and various ablation studies.

Abstract

We propose the structure and color based learned image codec (SLIC) in which the task of compression is split into that of luminance and chrominance. The deep learning model is built with a novel multi-scale architecture for Y and UV channels in the encoder, where the features from various stages are combined to obtain the latent representation. An autoregressive context model is employed for backward adaptation and a hyperprior block for forward adaptation. Various experiments are carried out to study and analyze the performance of the proposed model, and to compare it with other image codecs. We also illustrate the advantages of our method through the visualization of channel impulse responses, latent channels and various ablation studies. The model achieves Bjøntegaard delta bitrate gains of 7.5% and 4.66% in terms of MS-SSIM and CIEDE2000 metrics with respect to other state-of-the-art reference codecs.

SLIC: A Learned Image Codec Using Structure and Color

TL;DR

The structure and color based learned image codec (SLIC) is proposed in which the task of compression is split into that of luminance and chrominance and the advantages of the method are illustrated through the visualization of channel impulse responses, latent channels and various ablation studies.

Abstract

We propose the structure and color based learned image codec (SLIC) in which the task of compression is split into that of luminance and chrominance. The deep learning model is built with a novel multi-scale architecture for Y and UV channels in the encoder, where the features from various stages are combined to obtain the latent representation. An autoregressive context model is employed for backward adaptation and a hyperprior block for forward adaptation. Various experiments are carried out to study and analyze the performance of the proposed model, and to compare it with other image codecs. We also illustrate the advantages of our method through the visualization of channel impulse responses, latent channels and various ablation studies. The model achieves Bjøntegaard delta bitrate gains of 7.5% and 4.66% in terms of MS-SSIM and CIEDE2000 metrics with respect to other state-of-the-art reference codecs.
Paper Structure (14 sections, 2 equations, 7 figures, 2 tables)

This paper contains 14 sections, 2 equations, 7 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Structure and Color Based LIC
  3. Experiments and Results
  4. Ablation Study
  5. Conclusion
  6. References

Figures (7)

  • Figure 1: Network architecture of the proposed SLIC model. Q represents the quantizer, AE and AD indicate arithmetic encoder and decoder respectively.
  • Figure 2: RD curves compared with various codecs for the Kodak dataset.
  • Figure 3: Comparison of reconstructed image patches from SLIC and Cheng2020, compressed at a bitrate of around 0.3 bpp. (Best when viewed enlarged on a monitor.)
  • Figure 4: Latent visualization of proposed SLIC, Cheng2020cheng_learned_2020, and Hyperpriorballe2018variational models for the image kodim21.png. (Best when viewed enlarged on a monitor.)
  • Figure 5: Impulse responses of image ClassA_8bit_BIKE_2048x2560_8b_RGB.png.
  • ...and 2 more figures