ADCT: A NEW HIGH QUALITY DCT BASED CODER FOR LOSSY IMAGE COMPRESSION


ADCT in 2008 was state-of-the-art lossy image compression method provided the best ratio-distrortion performance in the scientific literature. ADCT is FREE for scientific use.

N.N. Ponomarenko, V.V. Lukin, K.Egiazarian, J. Astola, "ADCTC: a new high quality DCT based coder for lossy image compression", CD ROM Proceedings of LNLA, Switzerland, August 2008, 6 p.



A research version of the coder (only for Photoshop grayscale RAW images coding) and test images are given below:

adct.zip (50 Kb) - ADCT.
testset1.zip (3 Mb) - set of 512x512 grayscale test images (16 images).
testset2.zip (25 Mb) - set of 2288x1712 grayscale test images (8 images).



Comparison of the coders JPEG, JPEG2000, and ADCT
Comparison of the coders JPEG, JPEG2000, and ADCT for entire set of the test images 2288x1712 pixels (see the paper).

JPEG - JPEG with the uniform quantization and deblocking
JPEG2000 - D. Taubman, M. Marcellin, JPEG 2000: Image Compression Fundamentals, Standards and Practice. Boston: Kluwer, 2002



Visual comparison (bpp for all compared methods is equal to 0.25):
1 - Original image, 2 - JPEG image, 3 - JPEG2000 image, 4 - ADCT image, 5 - Partition scheme for ADCT

Use mouse and keyboard to switch between images.


Key words: efficient DCT based lossy image compression, image compression using partition schemes, advanced image compression, the best ratio-distortion performance, heuristic context modeling, adaptive compression, signs coding, deblocking, context-based entropy coding, state-of-the-art lossy image compression, discrete cosine transform.

Back to main page...

Last changed: 2018-12-15.