Cool-chic literature

There are several published papers describing and enhancing Cool-chic. Reading some of these papers might help understanding how Cool-chic works.

Used in this repo

This repository is based on several papers successively enhancing Cool-chic. Check-out the release history to see previous versions of Cool-chic.

2024

• May: Overfitted image coding at reduced complexity, Blard et al

  • Study the complexity-performance trade-off of Cool-chic & introduce a near real-time CPU only decoder. Corresponds to Cool-chic 3.2

  • git clone https://github.com/Orange-OpenSource/Cool-Chic.git

• February: Cool-chic video: Learned video coding with 800 parameters, Leguay et al

  • Corresponds to Cool-chic 3.1 inter-frame coding scheme.

  • git clone --branch v3.1 https://github.com/Orange-OpenSource/Cool-Chic.git

2023

• December: C3: High-performance and low-complexity neural compression from a single image or video, Kim et al

  • Some improvements related to image coding have been integrated in Cool-chic 3.0

  • git clone --branch v3.0 https://github.com/Orange-OpenSource/Cool-Chic.git

• July: Low-complexity Overfitted Neural Image Codec, Leguay et al.

  • This constitutes Cool-chic 2

  • git clone --branch v2.0 https://github.com/Orange-OpenSource/Cool-Chic.git

2022

• December: COOL-CHIC: Coordinate-based Low Complexity Hierarchical Image Codec, Ladune et al.

  • This constitutes Cool-chic 1

  • git clone --branch v1.0 https://github.com/Orange-OpenSource/Cool-Chic.git

Other papers

Some other papers have enhanced or adapted Cool-chic to particular use-cases but have not (yet?) been integrated in this repository.

2024

• April: Human–Machine Collaborative Image Compression Method Based on Implicit Neural Representations, Li et al.

  • Propose to adapt Cool-chic to accomodate simultaneous coding for human and machine.

• January: Fast Implicit Neural Representation Image Codec in Resource-limited Devices, Liu et al.

  • Faster Cool-chic decoder

• January: Cool-Chic: Perceptually Tuned Low Complexity Overfitted Image Coder, Ladune et al

  • Cool-chic-based submission to the CLIC24 challenge

2023

• December: Hybrid Implicit Neural Image Compression with Subpixel Context Model and Iterative Pruner, Li et al.

  • Speed-up cool-chic decoding by removing the auto-regressive probability model

• November: Enhanced Quantified Local Implicit Neural Representation for Image Compression, Zhang et al.

  • Propose a CNN-based post-filter (similar to Cool-chic 2.0) & a SGA-based quantization close to the softround mecanism introduced by C3

• September: Implicit Neural Multiple Description for DNA-based data storage, Le et al.

  • Adapt Cool-chic to encode/decode image on DNA instead of binary files

• June: INR-MDSQC: Implicit Neural Representation Multiple Description Scalar Quantization for robust image Coding, Le et al.

  • Add multiple description into Cool-chic to cope with noisy transmission channels