Vol 27, No 4 (2018)

A mathematical model of visual system of the human observing images on optoelectronic device display is proposed. The model allows to formalize in linear approximation the physiological processes of registration and transformation of halftone object images in the human visual system in the presence of external and internal noise, and to estimate the detection probability of objects visual images in the form of Johnson’s equivalent bar patterns. On the basis of proposed model mathematical expressions for the visual system contrast sensitivity function are derived. Experimental studies on the measurement of this characteristic are carried out. A satisfactory agreement between the calculated and experimental results indicates the adequacy of the proposed mathematical model for the accepted restrictions on its applicability.

The paper considers the use of convolutional neural networks for the concurrent recognition of the gender and age of a person by video records of his face. The emphasis is on the incorporation of the approach into mobile video analytics systems. We have investigated the fusion of decisions obtained during the processing of each video frame, including the use of the classifier committee based on Dempster-Shafer theory. We propose the novel age prediction method using the evaluation of the expectation of the most probable ages. We have compared existing neural-net models with a specially trained modification of the MobileNet convolution network with two outputs. The experimental results are given for such data collections as Kinect, IJB-A, Indian Movie and EmotiW. As compared with other conventional methods, our approach makes it possible to increase the age and gender recognition accuracy by 2–5% and 5–10% respectively.

With the development of internet, man-machine interaction has tended to be more important. Precise speech recognition has become an important means to achieve man-machine interaction. In this study, deep neural network model was used to enhance speech recognition performance. Feedforward fully connected deep neural network, time-delay neural network, convolutional neural network and feedforward sequence memory neural network were studied, and their speech recognition performance was studied by comparing their acoustic models. Moreover, the recognition performance of the model after adding different dimension human voice features was tested. The results showed that the performance of the speech recognition system could be improved effectively by using the deep neural network model, and the performance of feedforward sequence memory neural network was the best, followed by deep neural network, time-delay neural network and convolutional neural network. Different extraction features had different improvement effects on model performance. The performance of the model which was added with Fbank extraction features was superior to that added with Mel-frequency cepstrum coefficient (MFCC) extraction feature. The model performance improved after the addition of vocal characteristics. Different models had different vocal characteristic dimensions.

The results of the study of synovial fluid (SF) of patients with knee joint osteoarthritis (OA). The optical analysis of the SF samples, harvested during exploratory punctures of knee joints of patients suffering the osteoarthrosis in different stages using the standard method, was made. A certain component composition of surface of SF samples that differs for healthy people and the patients, having knee joint osteoarthrosis in different stages, is possible to identify as a result of analysis with the use of Raman spectroscopy method. The introduced optical coefficients help to estimate the spectral composition of surface of SF samples of the patients, suffering the early and late stages of ОА.