Filippov

pp. 87-100

pp. 320-329

The authors propose formal criteria of estimation of steganographic algorithms in the context of undetectability of hidden information by the opponent. A method of creating optimal detection methods is also presented in the article. The article also contains the results of the analysis of known and prospective stenographic algorithms operation with the usage of methods constructed in this manner.

Usage of bilateral filters for image smoothing allows of removing small-scale noises and details, and, at the same time, saving clearness of borders. The authors present the results of the comparative analysis for four well-known algorithms of bilateral filtering and give recommendations for their usage for image digital processing.

Motion blur is a common defect that leads to blurry images with inevitable loss of information. If motion blur is shift-invariant it can be modeled as a convolution of a latent image with shift kernel, where the kernel describes the trace of the sensor. In that case removing the motion blur from an image becomes a deconvolution operation. The authors present a fast blind deconvolution method that produces a de-blurred image in only a few seconds. High speed of this method results from accelerating the steps of both kernel and latent image estimation in the iterative de-blurring process.