Scientific and methodological foundations for optimal identification, recognition, and classification of single-cell medical objects have been developed using hybrid convolutional neural networks with recurrent learning, as well as image transformation and filtering mechanisms. Algorithms for training networks with various architectures have been investigated. Adapted neural network architectures with independent recurrent layers have been proposed. Mechanisms for detection, localization of selected areas with distorted points, and error correction using the Daubechies wavelet transform 4, 7 were implemented. Support vector machines and gradient optimization mechanisms were proposed for network training. A software package has been developed for visualization, recognition, and classification of images of medical objects based on the use of standard search, recognition, and classification tools, and CUDA parallel computing technology.
| Mualliflar | R. Safarov |
|---|---|
| Jurnal | Samarqand davlat universiteti ilmiy tadqiqotlar axborotnomasi |
| Nashr sanasi | 2026-07-01 |
| Jild | 1 |
| Son | 3 |
| Betlar | 111-119 |
| DOI | 10.59251/2181-3973.2025.v1.138.1.3949 |
DOI: 10.59251/2181-3973.2025.v1.138.1.3949 · Maqolaning asl sahifasi
This paper examines the problem of ensuring the security of information exchange based on the principles of quantum computing. Within the scope of the study, the stages of user authentication, secret-key generation via…
Raman spectra of pure ethyl acetoacetate were obtained at room temperature and atmospheric pressure. The experimental and computational Raman spectra were compared, and PED (Potential Energy Distribution) analyses were…
Modeling of the adsorption of CH, CH₂, and CH₃ fragments on fullerene C₆₀ was performed. Interatomic interactions were described using a hybrid potential scheme implemented in LAMMPS. The hydrocarbon fragment was…
With a high spectral resolution of 0.1 cm⁻¹, the (CH3)2O···HF complex's gas phase infrared absorption spectra were recorded at temperatures between 188 in the 3200–3800 cm⁻¹ range, which corresponds to the HF stretching…
A direct electrochemical method for the determination of sulfur-containing compounds in motor fuels using copper-modified screen-printed electrodes (Cu-SPCE) and square-wave voltammetry has been developed. The electrode…
In this work, a new complex compound based on 4-nitroftalic acid and ethylenediamine of composition [C8H3NO6]2−·(C2H10N2)2+ was synthesized for the first time. The complex compound was grown from solution by slow…
Addressing the challenges of ambiguity in veterinary medicine, this study proposes a novel approach to cattle disease diagnosis using neutrosophic sets and the Sugeno fuzzy model. The research develops specialized…
The article considers montmorillonite and laponite as promising layered silicates for the development of adsorbents and functional materials. The interest in these materials is attributed to their unique structure…
This study investigates the specific activities of natural (226Ra, 232Th, 40K) and technogenic (137Cs) radionuclides in soils and food products from selected regions of Uzbekistan using scintillation gamma spectrometry…
In this work, the inverse spectral problem method is applied to integrate an equation composed of a combination of nonlinear KdV(+1) and KdV(-1) equations with an integral source. An infinite system of Dubrovin…
Samarqand davlat universiteti ilmiy tadqiqotlar axborotnomasi — barcha maqolalar