Prediction of compression ratio in lossy compression of noisy images

A. Zemliachenko (1) , R. Kozhemiakin (1) , B. Vozel (2) , V. Lukin (1)

1 Kharkov National University
2 IETR - Institut d'Électronique et des Technologies du numéRique

Requirements to data transmission in communication lines or to image storage can be such that methods of lossy image compression have to be applied. However, for many techniques of lossy image compression it is not an easy task to provide a desired compression ratio. Besides, noise can be present in images and it is desirable to take its characteristics into account in lossy compression. One way to get around the arising shortcomings is to predict compression ratio and to manage (vary) it accordingly. This paper considers some ways to predict compression ratio for the case of additive white Gaussian noise often used in literature. Compression techniques based on discrete cosine transform are studied. Compression in the neighborhood of optimal operation point is paid main attention. Meanwhile, other variants are studied as well. © 2016 National University Lviv Polytechnic.

Mots clés: Compression ratio (machinery); Digital storage; Discrete cosine transforms; Forecasting; Gaussian noise (electronic); Image compression; White noise; Additive White Gaussian noise; Communication lines; Compression techniques; Lossy compressions; Lossy image compression; Noisy image; Optimal operation point; Audio signal processing

Domaines: Réseaux et télécommunications [cs.NI]; Electronique

A. Zemliachenko, R. Kozhemiakin, B. Vozel, V. Lukin. Prediction of compression ratio in lossy compression of noisy images. 13th International Conference on Modern Problems of Radio Engineering, Telecommunications and Computer Science, TCSET 2016, Feb 2016, Lviv, Ukraine. pp.693--697, ⟨10.1109/TCSET.2016.7452156⟩. ⟨hal-01368100⟩