The article presents the results of the adaptation of the hybrid HMM-DNN speech synthesis model for use in automated speaker recognition system for critical use (ASRSCU). In particular, the process of learning the HMM-DNN speech synthesis model with the estimation of the difference between the posterior probability distributions of all HMM states and the actual a posteriori probability distribution, calculated by DNN, and the use of semantic information in the speaker recognition process, has been improved. The features that are observed in the sequence of frames to which the input phonogram is divided describe this information. The obtained results allowed improving the efficiency of the textdependent speaker recognition when using ASRSCU in a noisy acoustic environment. The article formulated measures for the structural integration of the HMM-DNN component in ASRSCU and describes the practical aspects of this process. In particular, the choice of the type and the method of normalization of the vectors of basic informative features at the frame level was substantiated, the number of HMM states and GMM parameters were determined depending on the parameters of the chosen formation model, and the procedure for interpreting the recognition results was described. The paper formulates measures to optimize the learning process of the ASRSCU with the HMM-DNN component, which will be exploited in noisy environments.
An optical method for determine the quantity of water in milk using the visible optical radiation range is proposed. On the basis of theoretical and experimental studies of the water-milk solution spectral characteristics the proposed method mathematical model was created. The mathematical modeling of passing of the visible range optical radiation through a water-milk solution on certain thickness of the solution layer is carried out. As a result of the modeling, the dependence of the output voltage of the photo-receiver based on a pair of photodiode-operating amplifier from the relative mass fraction of milk in the water-milk solution and the wavelength of the optical radiation in the visible range is obtained.
The paper considers the fundamentals of the design theory of stochastic diagnostic type observers that are invariant to unknown inputs. It presents a description of identification and malfunctions localization problem when the noise is present. The article provides theoretical substantiation of the decomposition procedure of optimal stochastic observers with indefinite inputs and noise. Optimality criterions of such observer are defined. The paper reveals invariance of a differential signal in relation to indefinite, uncontrollable perturbations. Optionally, in a case, when a differential signal has a Gaussian distribution, the procedure of decision making about fault availability is offered.
Multistage integration of visual information in the brain allows people to respond quickly to most significant stimuli while preserving the ability to recognize small details in the image. Implementation of this principle in technical systems can lead to more efficient processing procedures. The multistage approach to image processing, described in this paper, comprises main types of cortical multistage convergence. One of these types occurs within each visual pathway and the other between the pathways. This approach maps input images into a flexible hierarchy which reflects the complexity of the image data. The procedures of temporal image decomposition and hierarchy formation are described in mathematical terms. The multistage system highlights spatial regularities, which are passed through a number of transformational levels to generate a coded representation of the image which encapsulates, in a computer manner, structure on different hierarchical levels in the image. At each processing stage a single output result is computed to allow a very quick response from the system. The result is represented as an activity pattern, which can be compared with previously computed patterns on the basis of the closest match.