Search

Iris-based identification is a well-known and widely used technology. However, its use in mobile devices is limited due to the algorithmic complexity of the existing methods and their low universality with respect to various capturing conditions. A method for finding the iris area by approximating its boundaries by two circles is proposed. The relative error of determining the circle parameters must not exceed 5%. Convolutional neural networks optimized with respect to the number of parameters are applied. The proposed method outperforms the classical (not based on neural networks) methods for solving this segmentation problem. The operation of the proposed method is checked on open iris image databases.

Iris segmentation is an irreplaceable stage of iris recognition pipeline. Its quality hugely affects overall accuracy. Previously when conditions were mild and controlled the task was solved by image processing techniques and rule based approaches. Nowadays widespread of biometric technologies has relaxed operation conditions for such systems demanding more flexible and robust solutions. Constantly increasing data and sensors availability created fertile field for growth of machine learning methods capable to cope with complex conditions. The latest contributions to iris segmentation were made on this surge by leveraging abundant data and modern machine learning algorithms. In spite of previously achieved great results this work addresses even more challenging conditions that allows iris recognition to be used in wide range of real life cases. Novel CNN architectures are proposed in this work. They were designed to combine the latest achievements in classification and semantic segmentation fields. FCN and SegNet architectures have been picked up as prototypes and were strengthened by residual blocks. This allowed to make lightweight networks that could be shipped on various embedded platforms to successfully operate under less controllable environmental conditions. The approach allowed to obtain 0.93 and 0.92 IoU on original and modified CASIA-Iris-Lamp datasets which is a significant improvement in comparison with the results achieved before.

Aseries of N-oxides of N-vinylpyrrolidone and 2-methyl-5-vinylpyridine copolymers with various degrees of oxidation were prepared via oxidation of the copolymers by peracetic acid. The N-oxidation kinetic rate constants were calculated using UV spectrophotometry. The temperature dependence of the limiting N-oxidation was determined. The molecular-weight characteristics of the N-oxidized copolymers were established using 13C NMR methods. An approach based on direct UV spectrophotometry was proposed for practical monitoring of the reaction completion (degree of conversion).

Photometric, spectral, and polarimetric observations of FG Sge during the stage of increasing activity in its R Coronae Borealis type variability are presented and analyzed. a weakening of the current level of rare transient states of maximum brightness is interpreted as a consequence of increasing optical thickness of a permanent circumstellar envelope. The brightness of the He1-5 nebula in the instrumental system in the Uc and V bands is 20m.69±0m.13 and 23m.01±0m.58 from a square arc second, respectively. Luminosity was not detected in the R, B, and Ic bands.

We present an improved analytical model as well as a new set of multi-wavelength observation of the polarization eclipse of the Wolf-Rayet binary V444 Cygni (WN5+O6). Comparing the model with the observations yields an estimate of the O and Wolf-Rayet star radii as well as of the Wolf-Rayet mass-loss rate. For the O star we find R = 8.5 R⊙ and for the Wolf-Rayet star R ⊙. This values are in agreement with those derived by Cherepashchuk et al. from the detailed analysis of multiwavelength light curves. For the Wolf-Rayet mass-loss rate we obtain M = 7.5 x 10-6M⊙/yr, which is compatible with the dynamical values obtained from the rate of orbital period increase and with the value of dM/dt determined from the orbital double-wave modulation in polarization, but is at least 3 times smaller than the values derived from free-free radio fluxes and modeling of infrared spectral lines. However, no allowance has been made in calculating the mass-loss rates for inhomogeneities, for which evidence is increasing in hot star winds. If the wind of the WR star V444 Cygni is found to be clumpy, the radio/IR mass-loss rates are likely to be overestimated because of their dependency on the square of density. In such case, these values would probably have to be significantly decreased, bringing them closer to the polarization estimates, for which clumpy winds are irrelevant, providing the electron scattering remains optically thin. Full version was published in: Ap. J., 410, 342-356 (1993).