The descendants of Al-Fargani

AUTOMATIC DETERMINATION OF THE DEFORMATION INDEX OF CAST IRON MICROSTRUCTURE GRAINS USING C++ AND OPENCV

Sardor Karimov, Ilxomjon Samandarov — Mon, 18 Aug 2025 00:00:00 +0300

This paper presents a C++ and OpenCV-based algorithm for the automatic determination of geometric parameters (area, perimeter, aspect ratio, eccentricity, shape factor) of cast iron microstructure grains and the calculation of the deformation index (DI). The proposed method includes Otsu thresholding, morphological cleaning, contour detection, and ellipse fitting. For each grain, the DI value was determined, followed by statistical analysis (mean, median, minimum, maximum) and distribution plotting. This approach can serve as an effective tool for predicting the mechanical properties of metals and optimizing heat treatment processes.

DEVELOPMENT OF A WEB PLATFORM FOR EARLY BREAST CANCER DIAGNOSIS BASED ON ARTIFICIAL INTELLIGENCE AND OPTIMIZATION ALGORITHMS

Xaydarov Sherali — Sat, 30 Aug 2025 00:00:00 +0300

This study addresses early breast cancer detection using an AI-based approach. A web platform, “Smart Diagnosis,” analyzes users’ medical parameters and survey data. Its diagnostic algorithm relies on machine learning, optimal feature selection, and classification error minimization. The system processes user data, generates diagnostic outcomes, and aids medical decision-making. Practical aspects, algorithm modeling, diagnostic accuracy, and web integration are examined. Results show that the system enables highly accurate early detection of breast cancer and offers practical value in modern medicine.

Algorithm for extracting renal vessels from abdominal images

Axmad Xashimov — Sat, 30 Aug 2025 00:00:00 +0300

This article is devoted to algorithms for distinguishing renal veins from vascular images based on human abdominal cavity images. The article proposes an algorithm for distinguishing veins by color and thickness from contrast-enhanced images.

CLASSIFICATION MODELS OF ARTIFICIAL INTELLIGENCE IN THE TRAINING OF PATHOLOGISTS

Farxod Meliev, У.Р. Шадиев, Ф.М. Мелиев, З.А. Маликов — Fri, 12 Sep 2025 00:00:00 +0300

The article discusses the implementation of modern classification models of artificial intelligence in the training of pathologists. The software package "GistoDiagnozUZ" has been developed, including modules for preprocessing, segmentation and classification of histological images. Experiments on open and proprietary datasets have shown an increase in the AUROC value to 0.93, a reduction in the time for slide analysis by almost half and a decrease in the cognitive load of pathologists. The results confirm that the integration of transfer learning and original CNN architectures improves diagnostic skills and can become the basis for standardized AI support in educational programs on pathomorphology.

RESEARCH ON CLASSIFICATION OF POTATO DISEASE LEAVES BASED ON XCEPTION

Hongzhi Zhang — Wed, 17 Sep 2025 00:00:00 +0300

Aiming at the difficulty in identifying potato early blight and late blight, this paper conducts a disease classification study based on Xception. Due to the similar initial symptoms of the two diseases, the traditional expert identification method has many limitations, while accurate identification is crucial for disease prevention and control. The research takes relevant leaves as the object, adopts the "Potato Disease Leaf Dataset v1", which is divided into training set, validation set and test set. The model is optimized through four groups of comparative experiments.

MODERN POSTQUANTUM CRYPTOGRAPHIC ALGORITHMS AND THEIR EFFICIENCY

Alisher Arabboyev, Bakhtiyor Abdurakhimov , Ilkhom Boykuziyev — Wed, 17 Sep 2025 00:00:00 +0300

The development of quantum computers has serious implications for cryptography. Most symmetric and asymmetric cryptographic algorithms are vulnerable to quantum algorithms. The Grover search algorithm improves the efficiency of finding the key in symmetric schemes such as AES and 3DES in square root of time. The security of asymmetric algorithms such as RSA, Diffie-Hellman, and Elliptic Curve Cryptosystems (ECC) is based on the mathematical difficulty of prime factorization and discrete logarithm problems. The best available classical algorithms take exponential time. The Shor factorization algorithm solves these problems in polynomial time.

METHODS AND MEANS OF ENSURING INFORMATION SECURITY IN DISTANCE LEARNING ASSESSMENT SYSTEMS

Abdulaziz Abduhalimov, Nizomiddin Ochilov — Wed, 17 Sep 2025 00:00:00 +0300

This article analyzes the most current methods and technological tools used to ensure data integrity, confidentiality, and availability in remote assessment environments. Based on the latest scientific literature, case studies from leading e-learning platforms (Coursera, ProctorU, ExamSoft, Moodle), and international cybersecurity standards (ISO/IEC 27001, NIST SP 800-171, GDPR), solutions such as biometric authentication, artificial intelligence control, blockchain-based certification, and “Zero Trust” architecture are studied. The study shows that while technological solutions are evolving, ethical, legal, and compliance challenges remain. The article concludes with practical recommendations for educational institutions, policymakers, and EdTech developers to implement layered, flexible, and human-centered security models.

DEVELOPMENT OF AN ALGORITHM FOR CONSTRUCTING FRACTAL EARTH SURFACE USING THE L-SYSTEM METHOD

Elyor Saidkulov, Z.E. Ibrohimova — Thu, 18 Sep 2025 00:00:00 +0300

Fractals are known to have extremely complex shapes, which are composed of unique objects that are important for understanding fractal ideas and real-life applications. This work analyzes the main methods for creating terrain, landscapes, and reliefs using algorithmic methods. In this scientific research work, many processes in nature are complex and uncertain, and it is difficult to model them directly. Therefore, algorithmic approaches allow us to transform this complexity into controllable and repeatable methods, taking into account the basic properties of nature (self-repetition, randomness, dynamics of processes).

CYBER ATTACKS AND THEIR METHODS OF EXECUTION

Dilshoda Qurbonaliyeva — Thu, 18 Sep 2025 00:00:00 +0300

The article analyzes the types of cyber attacks, methods of their implementation and protection measures in connection with the development of modern information and communication technologies. The main motives of cybercriminals, the goals of attacks and cybersecurity methods are covered in detail. The article presents the classification of cyber attacks based on the analysis of scientific literature and proposes strategies to protect against them. The purpose of this research is to develop effective defense measures in the field of cybersecurity and information systems protection.

DEVELOPMENT OF AN ANALYSIS SYSTEM USING CLUSTERING PROBLEMS OF CODE RECYCLE ATTACKS IN FUZZY INFORMATION ENVIRONMENTS

Alisherjon Ashuraliyev, Umidjon Xamrakulov — Fri, 03 Oct 2025 00:00:00 +0300

This research study addresses the issues of analyzing code-reuse attacks in uncertain information environments and their clustering. Clustering methods were applied to identify interrelated attack parameters. Special attention was given to the use of clustering techniques for detecting various characteristics of attacks and grouping them accordingly. The developed intelligent system enables rapid assessment of attacks and effective decision-making in real time.

ANALYSIS OF ATMOSPHERIC AIR POLLUTION CAUSED BY MOTOR VEHICLES

Farrux Muradov, Olimjon Kucharov, Go‘zal Samanova, Maftuna Abdiraxmonova — Fri, 03 Oct 2025 00:00:00 +0300

This study investigates the causes of atmospheric pollution from motor vehicles and explores methods for reducing the concentration of harmful substances in the air. Based on the analysis of data from various sources, the main drivers and consequences of vehicular air pollution were identified. Findings indicate that nearly 80% of total emissions originate from automobiles, etroleum derivatives, and other toxic compounds into the atmosphere. The continuous increase in the number of vehicles intensifies the accumulation of hazardous substances, which adversely affect human health by disrupting physiological functions. As a result, potential approaches to mitigate atmospheric pollutant concentrations have been proposed.

APPLICATION OF NEURAL NETWORKS IN CRYPTOANALYSIS ON THE EXAMPLE OF THE VIGENIER ALGORITHM

Mirzo-Ulugbek Davlatov, Orif Allanov, Bakhtiyor Turdibekov — Fri, 03 Oct 2025 00:00:00 +0300

This paper investigates the application of neural networks in cryptanalysis processes using the Vigenere cipher as a case study. Although the Vigenere cipher, one of the classical polyalphabetic substitution algorithms, was historically regarded as a strong cryptosystem, modern computational capabilities and artificial intelligence approaches help reveal its weaknesses. In this study, a neural network model was built using the PyTorch library, and experiments were conducted to reconstruct plaintext from ciphertext. The experimental results demonstrated that neural networks can learn statistical patterns inherent in the Vigenere cipher and are capable of partially automating the decryption process.

OPTIMIZATION OF HVAC ENERGY CONSUMPTION UNDER UNCERTAINTY OF EXTERNAL CONDITIONS: A PROBABILISTIC MODELING APPROACH

Mohira Rahimova, Lyudmila Varlamova — Fri, 03 Oct 2025 00:00:00 +0300

Abstract: Heating, Ventilation, and Air Conditioning (HVAC) systems account for about 40% of building energy use. Their performance is strongly affected by uncertain external factors such as weather and stochastic occupancy, limiting deterministic control strategies. This study proposes a probabilistic optimization framework for HVAC management, modeling occupancy as a stochastic process and weather via probability distributions. Indoor temperature and CO₂ dynamics are described by stochastic differential equations within a constrained optimization problem. A stochastic Model Predictive Control (MPC) with Monte Carlo sampling is applied, achieving 10–15% expected energy savings and improved comfort reliability, especially in educational and office buildings.

A COMPARATIVE STUDY OF ACTIVATION FUNCTIONS IN DEEP LEARNING MODELS

Abbaz Primbetov, Navruz Akbarov — Fri, 03 Oct 2025 00:00:00 +0300

Activation functions play a vital role in the training dynamics and generalization performance of deep learning models. This study presents a comparative analysis of ten widely used activation functions—ReLU, Sigmoid, Tanh, ELU, SELU, Softplus, Softsign, Swish, GELU, and a custom spline-based function—within a unified convolutional neural network (CNN) architecture. All models were trained and evaluated on the CIFAR-10 dataset under identical experimental settings, including fixed learning rate, batch size, number of epochs, and architecture configuration.

MATHEMATICAL MODEL OF THE PROCESS OF TRANSPORT AND DISPERSION OF POLLUTANTS IN THE ATMOSPHERE AND ITS NUMERICAL SOLUTION

Farrux Muradov, Olimjon Kucharov, Go‘zal Samanova, Maftuna Abdiraxmonova — Fri, 03 Oct 2025 00:00:00 +0300

This article addresses the problem of constructing numerical solutions of hydrodynamic equations in the form of nonlinear partial differential equations describing the three-dimensional velocity of atmospheric air masses. By applying high-order approximation with respect to temporal and spatial variables, implicit were developed, and their numerical solutions were analyzed. The proposed method ensures the condition of absolute stability, while the accuracy of the solution increases to a quadratic order. Using the mathematical framework implemented in dedicated software, the possibility has been established to assess and predict the spatio-temporal evolution of pollutants emitted from industrial sources in the atmospheric boundary layer.

ANALYSIS OF DIGITAL IMAGE PROCESSING METHODS AND SOFTWARE TOOLS FOR ECHOCARDIOGRAM IMAGES

Nigora Djurayeva, Mexriddin Raximov — Fri, 03 Oct 2025 00:00:00 +0300

In this study, digital image processing and a convolutional neural network (CNN) were integrated to automatically detect mitral stenosis based on 32 echocardiogram videos. By applying data augmentation, the training dataset was expanded, and the CNN architecture, consisting of four convolutional blocks, was trained for binary classification. The model achieved 92% accuracy and 0.26 loss on the test dataset. The results demonstrated that these approaches can serve as an effective auxiliary tool in clinical diagnosis.

Research Article Geometry-aware Analysis of Wheat Grain Hydration Prior to Milling

Abdushukur Urinboev, B Ismailov — Fri, 03 Oct 2025 00:00:00 +0300

This study analyzes geometry and short-time hydration of two wheat varieties from Uzbekistan, “Asr” and “Alekseevich.” Grain dimensions were measured by digital caliper, microstructure examined microscopically, and surface/volume reconstructed from 3D scans. Hydration kinetics were recorded gravimetrically at 5-min intervals for 30 min under 20–25 °C and 40–50 °C. “Asr” grains were larger and showed faster water uptake, with higher mass gain than “Alekseevich.” An ellipsoidal multi-layer diffusion model confirmed geometry-dependent moisture penetration. Results indicate that geometry-aware conditioning enhances millability and flour quality while minimizing repeated wetting–drying cycles.

DETECTION OF PRINTED FALSE ATTACKS USING NEURAL NETWORKS

Бахтиёр Абдукадиров — Fri, 03 Oct 2025 00:00:00 +0300

This article discusses a method for detecting false positives against a biometric facial recognition system based on deep convolutional neural networks. The proposed method is designed to detect printed false positives and is tested on open databases of real and fake faces, and the results are analyzed. The types of false positive attacks launched against a biometric system based on existing faces are analyzed.