Introduction. Federal industrial safety standards and regulations establish the requirements for the technical condition assessment of the said engineering devices and associated equipment. However, these regulatory frameworks require a discrete-level defect analysis of composite nodes and parts of load-lifting machinery. This approach does not always ensure the required accuracy and objectivity in monitoring the overall condition of equipment, as various combinations of defects in numerous structural components determine varying degrees of their overall wear. This is due to the fact that, during long-term operation, the components and parts of the mechanisms are subject to uneven destructive stress resulting in variability in the degree of their damage. The present work proposes a combination of scientific and technical measures to improve the operational reliability of both lifting machines and associated equipment through the integration of modern intelligent algorithms. The research objective of the study is to develop an intelligent decision support system designed for a comprehensive, integrated assessment of the operating conditions of lifting structures.

Materials and Methods. To achieve this objective, three models of artificial neural networks have been developed. The first model contains five layers: the first input layer includes eight neurons corresponding to the same number of rejection indicators of lifting structure components, such as defects in the undercarriage, braking system, rope-and-pulley system, and load hook. The three subsequent hidden layers contain eight neurons each. They are responsible for the learning process of the network and provide recording of intermediate results. The output layer consists of three neurons, each of which characterizes a certain class of technical condition of the object under study. The second neural network has a similar configuration, except for the input layer, which additionally includes nine neurons corresponding to the rejection indicators of the supporting metal structures of lifting facilities. The input layer of the third model contains twenty-nine neurons. This structure allows for a comprehensive assessment of equipment condition across a full range of rejection indicators. The practical implementation of the developed models was performed in the Python programming environment using the Scikit-learn machine learning library. Testing of the performance and efficiency of the developed networks was carried out on ten independent samples according to their performance criteria and confidence intervals for assessing the state of engineering devices.

Results. The performance of the developed models under testing was 100%, meaning each of the three neural network structures accurately classified the conditions of the lifting structures in all ten test scenarios. Moreover, the confidence level of the judgments made increased as the number of network input parameters was scaled in the range from 0.628 to 0.705.

Discussion. The presented data demonstrate the significant impact of the artificial neural network architecture on the precision of lifting structure diagnostics. The key result is the establishment of a numerical relationship between combinations of rejection criteria and the resulting condition class through the analysis of the activation levels of output layer neurons, which serve as a quantitative measure of the reliability of the decision made by the model. The findings correlate with previous research in the field of applying machine learning methods to technical diagnostics. The specificity of the proposed approach is in the integrated assessment of the overall condition without detailed mathematical modeling of local physical processes, which determines its efficiency under conditions of limited a priori information about operational modes. Limitations of the work include a relatively small sample size and a certain subjectivity of expert marking. Possible errors are due to the risk of retraining the neural network on small arrays of training information.

Conclusion. The proposed intelligent system is developed taking into account practical experience in operating lifting machines, analysis of accumulated statistical data, and the requirements of current industry standards. Scientific results expand the understanding of the potential of using artificial intelligence algorithms to ensure the technological safety of complex technical objects. The implementation of such systems will allow engineering and technical personnel who do not have extensive practical experience in on-site inspection of structures to make qualified and informed decisions regarding the possibility of continuing the safe operation of equipment.

Introduction. The recognition of basic emotions from facial images is increasingly required in driver monitoring systems, medical user interfaces, and educational learning analytics. As of July 2024, such systems have been mandated by EU regulations. Among the methodological approaches in this area, neural-network architectures reach high accuracy but require large training samples and remain opaque, which is critical in safety applications. Classical descriptors LBP and HOG, in contrast, are computationally efficient and interpretable. However, their performance under comparable conditions remains underexplored: there are still questions regarding the stability of their metrics under cross-validation on compact datasets, as well as the nature of confusions between visually similar emotions. In this context, it was assumed that the near-perfect accuracy of HOG observed in such protocols reflected leakage of subject identities between the training and test sets, thereby overestimating the generalization ability of the descriptor. The objective of the study was to evaluate LBP and HOG on the JAFFE dataset in a single, reproducible pipeline, and determine the limits of their applicability.

Materials and Methods. To address the stated tasks, the study was organized as a comparative empirical evaluation of two classical descriptors within a single reproducible protocol. The open JAFFE dataset (Lyons et al., 1998; 213 grayscale images of ten subjects with seven basic emotions) served as the experimental basis, with quality assessed under stratified five-fold cross-validation. The images underwent intensity normalization (histogram equalization) and geometric alignment by facial landmarks. Then, on the prepared frames, the LBP (texture) and HOG (contour geometry) descriptors were extracted, with specific parameters reported in the body of the article. The resulting features were fed to a linear SVM (support vector machine), with the regularization parameter tuned using a grid search over {0.01, 0.1, 1, 10} via internal three-fold cross-validation. Quality was assessed as the mean ± standard deviation of Accuracy and Macro-F1 metrics across five outer folds. The software implementation of the entire pipeline was done in Python 3.10 (using the scikit-learn 1.3 and scikit-image 0.21 libraries).

Results. In the experiment, HOG combined with linear SVM provided complete separation of the seven emotional classes on the JAFFE dataset within the selected stratified protocol. Under the same algorithmic order, LBP showed significantly lower accuracy and revealed a specific structure of systematic confusion errors: the pairs “fear — surprise” and “sadness — neutral” provided consistently indistinguishable, while the category-wise distribution of metrics quantified the fundamental differences between contour and texture feature representations.

Discussion. The results obtained indicate the decisive impact of the data splitting protocol on the final assessment of classical descriptors. Specifically, the perfect separability of HOG features is explained by preservation of subject identities across training and test sets, and is fully consistent with the known sensitivity of the gradient profile to individual facial features. With respect to LBP, the observed values fall within the range known from the reference work by Shan, Gong, and McOwan. However, the pattern of false positives documented in the present study is the first to be interpreted in detail through FACS description of overlapping sets of active facial muscles, which has previously been absent from published comparative reviews.

Conclusion. The work has successfully solved the problems of comparative evaluation of LBP and HOG methods, construction of a fully reproducible software pipeline, and meaningful interpretation of automatic classification errors. Based on the data obtained, the HOG model combined with a linear SVM is suitable for laboratory tasks with a fixed set of subjects, whereas LBP in the same combination is justified as a component of hybrid architectures with convolutional or recurrent networks. In applied terms, the results can be used in the design of emotion recognition systems for medical interfaces and driver-monitoring systems. At the same time, the limitations of the work include the compact JAFFE dataset and the laboratory imaging conditions. Addressing these limitations, future work will involve moving to a more demanding protocol with full subject-out exclusion from the training set, as well as a comparative evaluation against current deep learning models.