All issues

We suggested some approaches for solving image processing tasks in the decision support system (DSS) of the Center for Automated Recording of Administrative Offenses of the Road Traffic (CARAO). The main task of this system is to assist the operator in obtaining accurate information about the vehicle registration plate and the vehicle brand/model based on images obtained from the photo and video recording systems. We suggested the approach for vehicle registration plate recognition and brand/model classification on the images based on modern neural network models. LPRNet neural network model supplemented by Spatial Transformer Layer was used to recognize the vehicle registration plate. The ResNeXt-101-32x8d neural network model was used to classify for vehicle brand/model. We suggested the approach to construct the training set for the neural network of vehicle registration plate recognition. The approach is based on computer vision methods and machine learning algorithms. The SIFT algorithm was used to detect and describe local features on images with the vehicle registration plate. DBSCAN clustering was used to detect and delete outliers in such local features. The accuracy of vehicle registration plate recognition was 96% on the testing set. We suggested the approach to improve the efficiency of using the ResNeXt-101-32x8d model at additional training and classification stages. The approach is based on the new architecture of convolutional neural networks with “freezing” weight coefficients of convolutional layers, an additional convolutional layer for parallelizing the classification process, and a set of binary classifiers at the output. This approach significantly reduced the time of additional training of neural network when new vehicle brand/model classification was needed. The final accuracy of vehicle brand/model classification was 99% on the testing set. The proposed approaches were tested and implemented in the DSS of the CARAO of the Republic of Tatarstan.

The advent of laser scanning technologies has revolutionized forestry. Their use made it possible to switch from studying woodlands using manual measurements to computer analysis of stereo point images called point clouds.

Automatic calculation of some tree parameters (such as trunk diameter) using a point cloud requires the removal of foliage points. To perform this operation, a preliminary segmentation of the stereo image into the “foliage” and “trunk” classes is required. The solution to this problem often involves the use of machine learning methods.

One of the most popular classifiers used for segmentation of stereo images of trees is a random forest. This classifier is quite demanding on the amount of memory. At the same time, the size of the machine learning model can be critical if it needs to be sent by wire, which is required, for example, when performing distributed learning. In this paper, the goal is to find a classifier that would be less demanding in terms of memory, but at the same time would have comparable segmentation accuracy. The search is performed among classifiers such as logistic regression, naive Bayes classifier, and decision tree. In addition, a method for segmentation refinement performed by a decision tree using logistic regression is being investigated.

The experiments were conducted on data from the collection of the University of Heidelberg. The collection contains hand-marked stereo images of trees of various species, both coniferous and deciduous, typical of the forests of Central Europe.

It has been shown that classification using a decision tree, adjusted using logistic regression, is able to produce a result that is only slightly inferior to the result of a random forest in accuracy, while spending less time and RAM. The difference in balanced accuracy is no more than one percent on all the clouds considered, while the total size and inference time of the decision tree and logistic regression classifiers is an order of magnitude smaller than of the random forest classifier.

The article presents a technology for building clinical decision support systems (CDSS) based on service containers using Docker and a web interface that runs directly in the browser without installing specialized software on workstation of a clinician. A modular architecture is proposed in which each application module is packaged as an independent service container combining a lightweight web server, a user interface, and computational components for medical image processing. Communication between the browser and the server side is implemented via a persistent bidirectional WebSocket connection with binary message serialization (MessagePack), which provides low latency and efficient transfer of large data. For local storage of images and analysis of results, browser facilities (IndexedDB with the Dexie.js wrapper) are used to speed up repeated data access. Three-dimensional visualization and basic operations with DICOM data are implemented with Three.js and AMI.js: this toolchain supports the integration of interactive elements arising from the task context (annotations, landmarks, markers, 3D models) into volumetric medical images.

Server components and functional modules are assembled as a set of interacting containers managed by Docker. The paper discusses the choice of base images, approaches to minimizing containers down to runtime-only executables without external utilities, and the organization of multi-stage builds with a dedicated build container. It describes a hub service that launches application containers on user request, performs request proxying, manages sessions, and switches a container from shared to exclusive mode at the start of computations. Examples of application modules are provided (fractional flow reserve estimation, quantitative flow ratio computation, aortic valve closure modeling), along with the integration of a React-based interface with a three-dimensional scene, a versioning policy, automated reproducibility checks, and the deployment procedure on the target platform.

It is demonstrated that containerization ensures portability and reproducibility of the software environment, dependency isolation and scalability, while the browser-based interface provides accessibility, reduced infrastructure requirements, and interactive real-time visualization of medical data. Technical limitations are noted (dependence on versions of visualization libraries and data formats) together with practical mitigation measures.

Reliable plant disease diagnosis requires not only high classification accuracy, but also stable generalization and interpretable decision-making. Although attention mechanisms are useful to these deep learning models, the performance also depends on where and how they are integrated into the network architecture. This study presents a design-optimized YOLO11m-based classification framework that systematically investigates the impact of Convolutional Block Attention Module (CBAM) injection at different architectural levels for plant disease diagnosis. We perform a comparative modelcontrolled analysis of three model architectures: (i) the baseline YOLO11m-Cls architecture lacking attention, (ii) only adding the backbone block along with CBAM and (iii) a hybrid architecture that includes reduced backbone attention combined with CBAM added at classification head. All models are trained and tested under the same experimental settings using a largescale dataset with around 90 000 images of 38 types of plant diseases. Experimental results clearly show that the uniform injection of CBAM into the backbone reduces stability but causes generalization to worsen with a higher validation loss and significantly lower Top-1 accuracy (≈ 90.5%), while the hybrid attention design balances stability and discrimination, with Top-1 accuracy up to 99.71%, Top-5 accuracy up to 99.99% and near-baseline validation behavior respectively Grad-CAMbased interpretability analysis also demonstrates that the hybrid model generates enhanced and biologically interpretable activation maps, which are more disease-specific with less distraction from background. Notably, the aim of our work is not for achieving superior performance over all classifiers but instead only to provide design-level evidence on how placing attention modulates rigidity and interpretability in YOLO-based classification models. The results provide practical architectural considerations for building dependable and interpretable AI systems in agriculture.

The article shows the relevance of research work in the field of creating control systems for unmanned vehicles based on computer vision technologies. Computer vision tools are used to solve a large number of different tasks, including to determine the location of the car, detect obstacles, determine a suitable parking space. These tasks are resource intensive and have to be performed in real time. Therefore, it is important to develop effective models, methods and tools that ensure the achievement of the required time and accuracy for use in unmanned vehicle control systems. In this case, the choice of the image representation model is important. In this paper, we consider a model based on the wavelet transform, which makes it possible to form features characterizing the energy estimates of the image points and reflecting their significance from the point of view of the contribution to the overall image energy. To form a model of energy characteristics, a procedure is performed based on taking into account the dependencies between the wavelet coefficients of various levels and the application of heuristic adjustment factors for strengthening or weakening the influence of boundary and interior points. On the basis of the proposed model, it is possible to construct descriptions of images their characteristic features for isolating and analyzing, including for isolating contours, regions, and singular points. The effectiveness of the proposed approach to image analysis is due to the fact that the objects in question, such as road signs, road markings or car numbers that need to be detected and identified, are characterized by the relevant features. In addition, the use of wavelet transforms allows to perform the same basic operations to solve a set of tasks in onboard unmanned vehicle systems, including for tasks of primary processing, segmentation, description, recognition and compression of images. The such unified approach application will allow to reduce the time for performing all procedures and to reduce the requirements for computing resources of the on-board system of an unmanned vehicle.

In single-photon emission computed tomography (SPECT), patients with bone disorders receive a radiopharmaceutical (RP) that accumulates selectively in pathological lesions. Accurate quantification of RP uptake plays a critical role in disease staging, prognosis, and the development of personalized treatment strategies. Traditionally, the accuracy of quantitative assessment is evaluated through in vitro clinical trials using the standardized physical NEMA IEC phantom, which contains six spheres simulating lesions of various sizes. However, such experiments are limited by high costs and radiation exposure to researchers. This study proposes an alternative in silico approach based on numerical simulation using a digital twin of the NEMA IEC phantom. The computational framework allows for extensive testing under varying conditions without physical constraints. Analogous to clinical protocols, we calculated the recovery coefficient (RC_max), defined as the ratio of the maximum activity in a lesion to its known true value. The simulation settings were tailored to clinical SPECT/CT protocols involving ^99mTc for patients with bone-related diseases. For the first time, we systematically analyzed the impact of lesion-to-background ratios and post-reconstruction filtering on RC_max values. Numerical experiments revealed the presence of edge artifacts in reconstructed lesion images, consistent with those observed in both real NEMA IEC phantom studies and patient scans. These artifacts introduce instability into the iterative reconstruction process and lead to errors in activity quantification. Our results demonstrate that post-filtering helps suppress edge artifacts and stabilizes the solution. However, it also significantly underestimates activity in small lesions. To address this issue, we introduce post-reconstruction correction factors derived from our simulations to improve the accuracy of quantification in lesions smaller than 20 mm in diameter.

In modern cardiology, accurate assessment of the functional significance of coronary artery stenoses is a critical factor for selecting treatment strategies and making informed clinical decisions. This paper presents an automated algorithm for processing dynamic X-ray angiographic image sequences aimed at estimating blood flow velocity. This parameter serves as the basis for determining the Quantitative Flow Ratio (QFR), which acts as an effective noninvasive alternative to traditional invasive fractional flow reserve (FFR) measurements. The proposed methodology successfully overcomes classic challenges of angiographic analysis, such as vessel motion artifacts during the cardio-respiratory cycle, variable contrast opacification, and the geometric complexity of the vascular tree in two-dimensional projections.

The presented processing workflow includes several key stages. Initially, frame preprocessing is performed to suppress noise and filter out the anatomical background. Subsequently, segmentation is implemented using a Sato filter and Otsu thresholding, followed by skeletonization to extract vessel centerlines. Particular attention is paid to the algorithm for automated identification of bifurcation points and the filtration of artifactual intersections caused by vessel overlapping. To ensure data continuity, a temporal tracking method for the target segment based on template correlation is applied, which is especially important during phases with low contrast agent concentration. The mathematical core of the algorithm is based on solving a 1D inverse problem for the advection-diffusion equation, allowing for the recovery of blood flow velocity from temporal intensity curves.

As part of the study, a detailed validation of the method was conducted by comparing automated calculation results with manual expert measurements across ten clinical datasets. The results confirm the robustness of the computational scheme within physiologically relevant ranges and its ability to significantly reduce inter-observer variability. The developed approach minimizes the need for physician intervention in the data processing stage, opening up prospects for creating real-time clinical decision support systems in the catheterization laboratory setting.

We propose an approach for the reconstruction and quantitative phenotyping of plant morphological traits at early ontogenetic stages based on digital image analysis. The proposed algorithm combines deep learning and graph-based representations while incorporating biological principles of morphogenesis. This integration enables the transition from binary segmentation to the reconstruction of a topologically consistent plant structure with accurate separation of intersecting root systems in images containing multiple plants. At the first stage, binary masks of seeds, shoots, and root systems are generated using a U-Net convolutional neural network architecture. The resulting image is transformed into a graph model in which edges correspond to root and shoot segments, while vertices represent key morphological points, including branching and intersection nodes. A directed traversal algorithm initialized from the seed point, combined with watershed-based graph partitioning and a composite scoring function for primary axis selection, identifies individual plants as isolated subgraphs and accurately distinguishes the primary root, lateral roots, and shoot. The validity of the algorithm was confirmed through a multi-level validation procedure, including comparison of graph reconstruction results on segmented images and end-to-end evaluation of the complete computational pipeline on original images against existing software solutions and expert manual annotations. The proposed approach is robust to variability in root system morphology and image noise and provides high accuracy in morphological trait extraction. The results of this study can be applied in plant breeding programs.

This work analyzes the problem of traffic signs recognition in intelligent transport systems. The basic concepts of computer vision and image recognition tasks are considered. The most effective approach for solving the problem of analyzing and recognizing images now is the neural network method. Among all kinds of neural networks, the convolutional neural network has proven itself best. Activation functions such as Relu and SoftMax are used to solve the classification problem when recognizing traffic signs. This article proposes a technology for recognizing traffic signs. The choice of an approach for solving the problem based on a convolutional neural network due to the ability to effectively solve the problem of identifying essential features and classification. The initial data for the neural network model were prepared and a training sample was formed. The Google Colaboratory cloud service with the external libraries for deep learning TensorFlow and Keras was used as a platform for the intelligent system development. The convolutional part of the network is designed to highlight characteristic features in the image. The first layer includes 512 neurons with the Relu activation function. Then there is the Dropout layer, which is used to reduce the effect of overfitting the network. The output fully connected layer includes four neurons, which corresponds to the problem of recognizing four types of traffic signs. An intelligent traffic sign recognition system has been developed and tested. The used convolutional neural network included four stages of convolution and subsampling. Evaluation of the efficiency of the traffic sign recognition system using the three-block cross-validation method showed that the error of the neural network model is minimal, therefore, in most cases, new images will be recognized correctly. In addition, the model has no errors of the first kind, and the error of the second kind has a low value and only when the input image is very noisy.

Predicting novel drug properties is fundamental to polypharmacology, repositioning, and the study of biologically active substances during the preclinical phase. The use of machine learning, including deep learning methods, for the identification of drug – target interactions has gained increasing popularity in recent years.

The objective of this study was to develop a method for recognizing psychotropic effects and drug mechanisms of action (drug – target interactions) based on an analysis of the bioelectrical activity of the brain using artificial intelligence technologies.

Intracranial electroencephalographic (EEG) signals from rats were recorded (4 channels at a sampling frequency of 500 Hz) after the administration of psychotropic drugs (gabapentin, diazepam, carbamazepine, pregabalin, eslicarbazepine, phenazepam, arecoline, pentylenetetrazole, picrotoxin, pilocarpine, chloral hydrate). The signals were divided into 2-second epochs, then converted into $2000\times 4$ images and input into an autoencoder. The output of the bottleneck layer was subjected to classification and clustering using t-SNE, and then the distances between resulting clusters were calculated. As an alternative, an approach based on feature extraction with dimensionality reduction using principal component analysis and kernel support vector machine (kSVM) classification was used. Models were validated using 5-fold cross-validation.

The classification accuracy obtained for 11 drugs during cross-validation was $0.580 \pm 0.021$, which is significantly higher than the accuracy of the random classifier $(0.091 \pm 0.045, p < 0.0001)$ and the kSVM $(0.441 \pm 0.035, p < 0.05)$. t-SNE maps were generated from the bottleneck parameters of intracranial EEG signals. The relative proximity of the signal clusters in the parametric space was assessed.

The present study introduces an original method for biopotential-mediated prediction of effects and mechanism of action (drug – target interaction). This method employs convolutional neural networks in conjunction with a modified selective parameter reduction algorithm. Post-treatment EEGs were compressed into a unified parameter space. Using a neural network classifier and clustering, we were able to recognize the patterns of neuronal response to the administration of various psychotropic drugs.