Lossless Compression

Update 20 July 2012: Binaries for lossless audio compression added.
Update 17 May 2012: New paper added.
Update 3 April 2012: Natural grayscale test images added.

Lossless Image Compression

Introduction
Medical Image Dataset
Natural Image Dataset
Binary and Matlab files
Acknowledgment
Contact
Related Papers
Links

Introduction

Lossless image compression is important in cases of when data acquisition is expensive, due to legal constrains or due to possible loss of valuable information. Lossless compression of medical images is especially important to preserve diagnostic value of images or to enable quality lossless transformation.

Medical Image Dataset

Dataset consists of 1338 previously uncompressed 512x512 grayscale CT images, with various statistical features. At this moment, only few images are available for testing.

Sample of medical images

All images are in pgm P5 format. The images provided here are for research purposes only.

Download images (841 kB)

Natural Image Dataset

Dataset consists of 147 previously uncompressed 680x1024 12-bit grayscale natural images and 154 16-bit grayscale natural images. All images are taken in Banja Luka, Bosnia and Herzegovina, usually in parks and around Vrbas river.

Sample of natural images

All images are in pgm P5 format. The images provided here are for research purposes only.

Download 12-bit images (144 MB)

Download 16-bit images (182 MB)

Binary executable of the proposed lossless image compression algorithm

Executable of proposed algorithm can be downloaded from this link. This file can be used for research purposes only.
Examples:
ImageCompression input.pgm output
ImageDecompression output newinput.pgm
ImageCompression input.pgm output -p 1 -t 8 12 20 (user defined predictor and thersholds)
Futher info can be seen with ImageCompression - help command

MATLAB simulation files for coding and decoding can be downloaded from this link.

Please report bugs on: aleksej@etfbl.net.

Acknowledgment

This research is related to the project called Lossless Image Compression, which is supported by the Ministry of Science and Technology of the Republic of Srpska, no. of contract 06/0-020/961-166/09. Also, authors would like to thank KBC Paprikovac, Banja Luka, who supplied medical images for testing that are used in this research.

Contact

E-mail: aleksej@etfbl.net

Related papers

A. Avramović, G. Banjac: "On Predictive-Based Lossless Compression of Images with Higher Bit Depths", Telfor Journal (accepted). (pdf)

A. Avramović: "Lossless Compression of Medical Images Based on Gradient Edge Detection", In Proceedings of 19th Telecommunications Forum TELFOR 2011, Belgrade, Serbia, 2011. (pdf)

A. Avramović: "Predictive-Based Lossless Compression of Medical Images", In Proceedings of 20th International Electrotechnical and Computer Science Conference ERK, Portorož, Slovenia, 2011. Vol. B 167-170.

A. Avramović, S. Savić: "Lossless Predictive Compression of Medical Images", Serbian Journal of Electrical Engineering, Volume 8. No. 1, pp. 27-36, February 2011. (Award for the best paper presented in Section Electric Circuits and Systems and Signal Processing, at Conference ETRAN 2010.) (pdf)

A. Avramović, B. Reljin: "Gradient Edge Detection Predictor for Image Lossless Compression" In Proceedings of 52th International Symposium ELMAR-2010, Zadar, Croatia, 2010. pp. 131-134. (pdf)

Links


Lossless Audio Compression

Introduction
Binary and Matlab files
Contact

Introduction

During the music recordings, mixing and production phases, it is important to store audio data without any losses. Lossless audio compression may be preferred in systems for voice recognition and similar applications. This application compress Wave files using simple prediction and modular arithmetic.

Binary executable of the proposed lossless audio compression algorithm

Executable of proposed algorithm can be downloaded from this file. This file can be used for research purposes only.
Examples:
MALACcoder input.wav output.mal
MALACdecoder output.mal newinput.wav
For help type MALACcoder -help command.

Contact

E-mail: aleksej@etfbl.net