Vladislav V. Sergeyev

Compression method for real-time systems of remote sensing

In this paper the method of image compression is presented. It is designed for data processing in real-time systems of remote sensing. In the midpoint there are compression algorithm based on component transformation with pixel interpolation and algorithm of stabilization of encoded image forming speed, which provide high compression ratio, stable speed of an output data flow and controlled error of image reconstruction.

Application of polynomial bases for image processing using sliding window

The work deals with the application of polynomial bases for digital image processing using sliding window. An algorithm is built for the parallel-recursive calculation of local moment characteristics. A parametric set of polynomial bases is introduced that yields the fastest realization of the algorithm. We consider methods of the calculation of the polynomial approximation parameters for the convolution kernel. The examples are adduced of the employment of polynomial bases for the 2-D signal filtration, and for the detection and recognition of objects on the image.

Detection of objects on the image using a sliding window mode

A procedure is proposed for processing images and detecting objects using a sliding window mode and allowing for an effective realization in terms of the computational complexity and the quality of detection. The main stages of data transformation are reported: preliminary image processing, recursive calculation of local features, the generation of the field of values for the discriminant function and object localization. The algorithm of the parametric set-up is developed and realized in the form of the learning of a linear classifier. An example is presented that shows the efficiency of the developed detection procedure.

Gabor Filter Based Attack on Printed Documents Protection Methods via Digital Watermarks

In this paper an attack on information hiding algorithms used for protection of printed documents or covert communication through them is proposed. This universal attack is applicable to a broad class of algorithms, in which information embedding is carried out by visually imperceptible distortion of background texture of documents. This attack, based on spectral analysis of electronic copies of printed documents and used Gabor filtering, makes it possible to extract the embedded information, and in some cases to determine the key of the attacked watermarking or steganography system. This key can then be used for insertion of own information in any appropriate container. Examples of successful data extraction using the proposed technique are shown.

Extraction of latent images from printed media

In this paper we propose an automatic technology for extraction of latent images from printed media such as documents, banknotes, financial securities, etc. This technology includes image processing by adaptively constructed Gabor filter bank for obtaining feature images, as well as subsequent stages of feature selection, grouping and multicomponent segmentation. The main advantage of the proposed technique is versatility: it allows to extract latent images made by different texture variations. Experimental results showing performance of the method over another known system for latent image extraction are given.

Identification of model parameters and correcting filters for space-variant distortions

We consider topics related to the construction of reconstructing filters for the correction of nonisoplanar distortions of defocusing type. The problem is tackled by the direct identification of shift-invariant models and reconstructing filters on small image fragments. We propose and study a new method for the selection of fragments based on analyzing the conjugation of vectors of independent variables with zero-space.

Parallel-recursive local image processing and polynomial bases

For image processing using sliding window mode, it is shown that the algorithms realizing the parallel-recursive calculation of the convolution with the approximation of the impulse response FIR-filter by polynomial bases are the best suited.

Stabilization of the compressed data formation rate in hierarchical image compression

A new method is proposed for stabilizing the rate of compressed data formation in the case of hierarchical image compression. The method is based on using various values of the control parameter (maximum error) for various scale levels of image representation and for error correction at the last level depending on the received compressed information content.

Using hierarchical histogram representation for the EM clustering algorithm enhancement.

This paper is devoted to EM clustering improvement using hierarchical multivariate histogram for probability density representation. We propose to store and operate with the image histogram by means of a special tree data structure. This allows to speed up computations in the case of multivariate input. We also answer the questions of the algorithm initialization and offer an initialization rule, which exploits the proposed histogram-tree structure. We have tested our algorithm modification and initialization rule using remote sensing images. Obtained results have confirmed that the modified algorithm is faster and the initialization rule provides better clustering in comparison with the traditional EM algorithm implementation.

The Algorithms of Hierarchical Histogram computation for multichannel images

This article is devoted to a simple approach to overcoming dimensionality problem in the case of multichannel image histogram computation. It is an extremely important issue for big multispectral and hyperspectral images having dozens and hundreds of image components. We propose a simple technique to compute and to store multivariate histograms based on the binary representation of pixel vector. Computed histograms can be further used in other image processing algorithms. In contrast to the other density estimation methods for multivariate data, our approach allows to compute histogram much faster. It also more memory effective than traditional linear (list or table) structured histograms. We provide some theoretical explanations on memory and time expenses for the proposed algorithms and traditional linear (list or table) structured histograms and describe the results of experimental evaluation of the algorithms. Experimental research has been produced with the set of multispectral remote sensing images.