Prediction of compression ratio in lossy compression of noisy images
Alexander Zemliachenko, Ruslan A. Kozhemiakin, Benoît Vozel, Владимир Васильевич Лукин · 2016
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.