This repository contains the source code resulting from the development of an ANS based coder for JPEG-LS.
In this work we propose enhancements to the JPEG-LS standard, aimed at improving its coding efficiency at a low computational overhead, particularly for hardware implementations. The main contribution is a low complexity and efficient coder, based on Tabled Asymmetric Numeral Systems (tANS), well suited for a wide range of entropy sources and with a simple hardware implementation. This coder enables further optimizations, resulting in great compression ratio improvements. When targeting photographic images,the proposed system is capable of achieving, in mean, 1.6%, 6% and 37.6% better compression for error tolerances of 0, 1 and 10, respectively. Additional improvements are achieved increasing the context size and image tiling, obtaining 2.3% lower bpp for lossless compression. Our results also show that LOCO-ANS compares favorably against state-of-the-art codecs like JPEG-XL and WebP, particularly in near-lossless, where it achieves higher compression ratios with a faster coding speed.
This work has an associated publication (currently under revision):
Title: "LOCO-ANS: An optimization of JPEG-LS using an efficient and low complexity coder based on ANS"
Authors: Tobías Alonso, Gustavo Sutter, and Jorge E. López de Vergara
Submitted to IEEE Access
- codec: source code of LOCO-ANS codec
- notebooks:
- coder_config_gen.ipynb: Jupyter notebook to select coder parameters and generate the configuration file and tANS tables
- results_images
- Test: test code
- tools: auxiliary tools
A quick test can be run using the BASH script under the Test folder.
To run the test, the codec and auxiliary tools need to be built (run make in the repo root)
Prerequisites:
Dataset: Rawzor Gray 8 bit images where used
Complete dataset:
Photographic images of the dataset:
Dataset: CLIC images subset
Average encoder MiPixels/s versus average bpp for software implementations of different compression algorithms. Numbers next to each point indicate the corresponding peak error. Pareto frontier is drawn with a solid line for error tolerances in [0,1,2]
Average decoder MiPixels/s versus average bpp for software implementations of different compression algorithms. Numbers next to each point indicate the corresponding peak error. Pareto frontier is drawn with a solid line for error tolerances in [0,1,2]