This paper is devoted to the problem of corner detection in manifold-valued images and in vector fields on manifolds. Our solution is a generalization of the Harris corner detector (C. Harris, 1988). As in the grayscale case, our algorithm is based on an estimation of a self-similarity of a point neighborhood. We define the self-similarity for the general cases and obtain approximations of it by an action of a bilinear form. This form can be viewed as a generalization of the structure tensor (M. Kass, 1987). The generalized structure tensor is then used as usual in the corner detection procedure. Finally, we describe future experiments: the algorithm will be tested on a task of chemical compounds classification.

This paper is devoted to the problem of blob detection in manifold-valued images. Our solution is based on new definitions of blob response functions. We define the blob response functions by means of curvatures of an image graph, considered as a submanifold. We call the proposed framework Riemannian blob detection. We prove that our approach can be viewed as a generalization of the grayscale blob detection technique. An expression of the Riemannian blob response functions through the image Hessian is derived. We provide experiments for the case of vector-valued images on 2D surfaces: the proposed framework is tested on the task of chemical compounds classification.

Prediction of the properties of chemical compounds by mathematical methods of pattern recognition is considered. The investigation was carried out by the example of the activity of cell division enzyme inhibitors. An approach based on mixtures of algorithms is used as the method for the construction of recognition models. A two-phase solution procedure for the structure–property problem is analyzed. The local classifier based on the nearest neighbor algorithm and the method of clustering sets is also described. New algorithms for the construction of classifier mixtures are compared. The methods of coordinated prediction of the activity of new compounds are examined. A comparison of mathematical modeling results with molecular design methods based on the coordination of compounds with known structures of therapeutic targets is also presented. An experimental study of the biological activity is conducted.

3D-QSAR and molecular docking were applied to predict the inhibitory activity of 196 compounds towards poly(ADP-riboso) polymerase-1 (PARP). A proportion of experimentally active ligands was higher among compounds with good rankings from both methods (57%) compared to compounds scored as inactive by at least one method (40% for docking-active, QSAR-inactive compounds).

A new approach to analysis of the molecule-descriptor matrix in the structure-property problem, based on the fuzzy cluster structure of the training sample, is developed. Methods for constructing fast prediction rejection rules and for the search of outliers in a training sample are described. To that end, a special space of easily computed descriptors is introduced. Optimization of the classifying function with respect to the parameters of fuzzy classification is considered. Prognostic models with a high quality of prediction, based on this approach, are proposed. Comparison of models is performed, which shows the efficiency of the described methods.

A new approach for analyzing the "molecule-descriptor" matrix for the QSAR problem (Quantitative Structure-Activity Relationship) based on a fuzzy cluster structure of the learning sample is presented. The ways for generating fast rules for refusing prediction and searching the spikes in the learning sample are described. For this purpose, a special space of descriptors, simple for calculation, is introduced. The ways for optimizing the discriminant function according to fuzzy clustering parameters are examined. Highly predictive models based on the presented approach have been generated. The models are compared, and the efficiency of the described methods is revealed.

A new approach for analyzing the moleculedescriptor matrix for the QSAR problem (Quantitative StructureActivity Relationship) based on a fuzzy cluster structure of the learning sample is presented. The ways for generating fast rules for refusing prediction and searching the spikes in the learning sample are described. For this purpose, a special space of descriptors, simple for calculation, is introduced. The ways for optimizing the discriminant function according to fuzzy clustering parameters are examined. Highly predictive models based on the presented approach have been generated. The models are compared, and the efficiency of the described methods is revealed.

The solution of the "structure-property" based on the molecular graphs descriptors selection with k-NN classifier is proposed. The results of comparing the construction of predictive models using the search and without it are given. The stability of the classifier function construction quality is tested using the test sample.

A method for solving the “structure-property” problem is proposed based on the adaptive choice of molecule description and automatic selection of a feature space according to the characteristics of the learning sample. A problem of combinatory explosion is solved using the method for group account of arguments. The algorithms for cluster analysis are used to improve the predictive capability of the model. The results of model construction using specific learning samples of chemical compounds are discussed.

A way to solve the QSAR problem (Quantitative Structure-Activity Relationship) by selecting the molecular graph descriptors using the k-NN classifier is presented. The predictive models by using the search and without it are generated and compared, and the results of comparison are presented. The stability of the discriminant function quality of generation is tested by using the test sample.

The report focuses on relative characteristics of the process management and activity management used in software development methodologies. As an example, the methodology IBM Rational Unified Process (RUP) and the Agile methodology are discussed. The use of the process management allows you to turn “hard” RUP methodology (with proper adaptation) in the Agile-RUP. Use of activity Management in Agile-projects significantly increases the risk of project failure in general, and “contradicts” the essence of the Agile methodology. The report reveals the characteristics of the process approach to management as an approach based on the quality in the broad sense. Process management took shape and grew up in such production organization methodologies as TQM (Total Quality Management), JIT (Just in Time), Six Sigma. An illustration of the process management characteristics are used as an example of best practices and techniques of methodology Agile.

A two-phase method for the pattern-driven recognition of objects in images is presented, implemented, and tested numerically. The method is based on the use of an active sensor. Possibilities for development are envisaged. This approach was shown to have advantages in solving the object-background separation problem and a high recognition rate was achieved with slow learning.

In the paper, a computational model for recognition of objects in a scene image is presented. The model is based on the use of an active sensor. The structure of the object model (OM) is described. This structure is a component that stores different representations of the object and puts at user’s disposal an interface whose operations are used in the scene recognition process.

A special algorithm for solving the QSAR problem for amber odorants has been considered.

A 3D-QSAR approach based on the electrostatic surface of molecules was used for the ambergris odour, and it showed a cross validation coefficient of 0.8.