Research and Production Complex Technological Centre

This article introduces MinMAE, a novel activation calibration method for Post-Training Quantization (PTQ) that significantly reduces accuracy loss in Convolutional Neural Networks (CNN). Motivated by the need for high-fidelity quantization without costly retraining, MinMAE directly minimizes the Mean Absolute Error (MAE) between original and dequantized activations, making it robust to outliers that degrade standard methods. We evaluated MinMAE against absmax and mean+3std on YOLOv8n and YOLOv8s models. For 8-bit quantization on YOLOv8n, MinMAE achieved 8.4% mean Average Precision (mAP) loss, whereas standard methods led to drops of up to 14.4%. On the larger YOLOv8s model, MinMAE also demonstrated superior performance, reducing loss from 7.2% to just 5.2%. While this approach requires a higher one-time calibration time, the substantial accuracy gains establish MinMAE as a highly effective and broadly applicable strategy for improving the stability and performance of quantized CNNs.

Results of the Rietveld X-ray diffraction analysis are presented for PbZr0.53Ti0.47O3, Pb0.95Sr0.05Zr0.53Ti0.47O3, PZT-19, and PZT-23 samples along with the results of studying their dielectric, piezo-, and pyroelectric properties. The samples are found to consist of a mixture of solid solutions having the perovskite structure with tetragonal (T), monoclinic (M), and rhombohedral (R) crystal lattices. Their weight fractions are 48T/39M/13R, 39T/61M, 49T/51M, and 72T/28M in the samples respectively. Main structural parameters of the phases are determined (unit cell symmetry and parameters of the phases, atomic coordinates and isotropic thermal displacements). The effect of the phase compositions of the samples on their electrophysical properties is discussed. Based on the results of the theoretical-group analysis, the conclusion is drawn that the observed differences in the ferroelectric and related (dielectric, piezoelectric) properties of the studied samples are due to: 1) concentration of monoclinic phase Cm, 2) structural parameters of the Cm phase.

Abstract Chrysolina (Ambrostomima) gusakovi sp. nov. and Ch. (Ambrostomima) katanovae sp. nov. are described from Hengduan Mountains (Sichuan and Gansu Provinces, respectively). They are externally close to Ch. amica, Ch. baimana, Ch. ferdinandoi, Ch. lijieae, Ch. viridiopaca, and Ch. wangi and differ mostly by the relief of pronotum and aedeagus structure. Chrysolina (Ambrostomima) amica species group including the eight above-mentioned species is proposed as part of the subgenus Ambrostomima.Chrysolina (Ambrostomima) angustipes sp. nov. is described from Hengduan Mountains (Sichuan Province). It is externally close to Ch. latipes and Ch. jinxiaoae and differs mostly in pronotal and elytral relief as well as aedeagus structure. Chrysolina (Ambrostomima) jinxiaoae species group including the three above-mentioned species is proposed as part of the subgenus Ambrostomima.Chrysolina (Ambrostomima) elliptica sp. nov. and Ch. (Ambrostomima) tristis sp. nov. are described from Gansu Province. They are externally close to Ch. korotjaevi and Ch. nigrorugosa and differs mostly in the aedeagus structure. Chrysolina nigrorugosa is transferred from the subgenus Chrysolina (Timarchomela) to the subgenus Ch. (Ambrostomima). Chrysolina (Ambrostomima) nigrorugosa species group including the four above-mentioned species is proposed as part of the subgenus Ambrostomima.Chrysolina (Ambrostomima) sarli sp. nov. is described from Hengduan Mountains (Sichuan Province). Externally it resembles Ch. aurita and differs mostly in pronotum relief as well as aedeagus structure. However, this new species does not form a group of morphologically close taxa with any other species within the subgenus Ch. (Ambrostomima).Chrysolina (Chrysocrosita) zoiai sp. nov. is described from Gansu Province. This species is morphologically close to Ch. alaschanica, Ch. fuyunica, Ch. josefi, Ch. nikolskyi, Ch. sogdiorum, Ch. spectabilis, and Ch. sulcicollis. Chrysolina (Chrysocrosita) spectabilis species group including the eight above-mentioned species is proposed as a part of the subgenus Chrysocrosita. For the first time, diagnostic differences in the structure of flagellum of aedeagus have been studied for the present subgenus. Based on the data on the ranges of species, an assumption has been made about the relict nature of the Ch. (Chrysocrosita) spectabilis species group.Chrysolina (Timarchomela) darwini sp. nov. and Ch. (Timarchomela) lamarcki sp. nov. are described from Hengduan Mountains (NW Yunnan Province). They are morphologically close to Ch. dalia, Ch. lii, and Ch. nushana and differ mostly by the aedeagus structure. Chrysolina (Timarchomela) dalia species group including the four above-mentioned species is proposed as part of the subgenus Timarchomela.Chrysolina (Volosatik) multipunctata sp. nov. is described from Hengduan Mountains (NW Yunnan Province). It is morphologically close to Ch. igori and differs mostly in pronotum relief and elytral puncturation in both sexes, as well as in pro- and meso-tarsomeres in male.

The contact printing method is widely used to create microchips and develop highly sensitive biosensors. This method is based on techniques for immobilizing biomolecules. In this work, two effective methods for covalent immobilization of oligonucleotides based on reactions of disulfide bond formation and azide–alkyne cycloaddition are optimized for the contact printing method. With both immobilization strategies, oligonucleotides retain the ability to hybridize with complementary DNA, which provides a basis for their practical use in bioanalytical applications.

Nanocrystalline TiO2 is a perspective semiconductor gas-sensing material due to its long-term stability of performance, but it is limited in application because of high electrical resistance. In this paper, a gas-sensing nanocomposite material with p-p heterojunction is introduced based on p-conducting Cr-doped TiO2 in combination with p-conducting Cr2O3. Materials were synthesized via a single-step flame spray pyrolysis (FSP) technique and comprehensively studied by X-ray diffraction (XRD), Brunauer-Emmett-Teller (BET) specific surface area analysis, transition electron microscopy (TEM), energy dispersive X-ray (EDX) spectroscopy, X-ray photoelectron spectroscopy (XPS), electron paramagnetic resonance (EPR), and Raman spectroscopy. Gas sensor performance in direct current (DC) mode was studied toward a number of gasses (H2, CO, CH4, NO2, H2S, NH3) as well as volatile organic compounds (VOCs) (acetone, methanol, and formaldehyde) in dry and humid conditions. The long-term stability of the obtained materials' gas sensor performance was evaluated alongside with an ex situ study of structural evolution. High sensitivity toward oxygenated VOCs and a lower detection limit below ppm level with a limited influence of humidity were shown. The long-term gas sensor performance stability of the obtained materials and its connection to the defect structure of doped TiO2 is demonstrated.

The work examines the problem of improving the quality of data obtained in the IR range, their primary processing for possible analysis by the operator. To perform these operations, the work considers an approach based on the use of a group of methods that are unified for a wide range of tasks and, as a consequence, allow processing on computationally simple devices. We also propose iterative primary processing algorithms based on the multi-criteria smoothing method and the extraction of local features on two-dimensional data. The estimated number of elementary operations required for its implementation is shown, processing examples are given, parameters and errors are errors. An algorithm for localizing closed areas according to the temperature threshold class is presented. An algorithm has been developed for parallel analysis of streaming information from a pair of IR sensors that generate data in different ranges. An approach to parallel data extraction and the formation of a new information field with combined data is considered. Using the example of field test data, the results of improving their quality and examples of the formation of combined IR data are shown.

Image segmentation is the critical step in different imaging and especially optical inspection applications: detection and recognition of objects, classification, analysis, and identification. Also, image gradient, as a preprocessing step, is an essential tool in image processing in many research areas, such as edge detection, segmentation, inpainting, etc. However, these tools have limitations and could be more accurate since the capture devices usually generate low-resolution images, which are primarily noisy and blurry. It is critical to receive useful gradient estimation on noisy color images while preserving the sharp edges. In the present paper, we develop a new gradient by integrating the quaternion framework with local polynomial approximation and the intersection of confidence intervals based on anisotropic gradient concepts for color image processing applications. We apply the proposed gradient technique in a modified active contour method to perform an automated segmentation for optical inspection applications. Computer simulations on the segmentation dataset for optical inspection applications show that the new adaptive quaternion anisotropic gradient exhibits fewer color artefacts than state-of-the-art techniques.

A switching platform that provides writing and erasing of extended two-phase periodic laser-induced surface structures on the GST has been developed: writing is carried out by fs laser pulses, and erasing - by electric current ones. The obtained results demonstrated the possibility of using a thin-film resistive heating element of the developed design to create elements of non-volatile active integrated optical and optoelectronic devices, for example, tunable diffraction gratings.