This PDF file contains the front matter associated with SPIE Proceedings Volume 13400, including the Title Page, Copyright information, Table of Contents, and Conference Committee information.

Wilga 2024 Summer Symposium on Photonics Applications and Web Engineering was the 52 th edition of the research and technical meetings series. Traditionally, the annual series of technical conferences and topical sessions lasted the whole week. They usually were split to one winter meeting in January and one or two summer events in May/June, and in September. The Covid epidemic and the war in Ukraine has changed a lot. The hybrid meetings were held in Wilga resort near Warsaw, owned by the Warsaw University of Technology. Now the meetings since 2022 were shifted to Lublin. Nearly 100 participants took part in all of the Wilga 2022 events, most of them being young researchers active in all aspects of photonics, electronics and ICT science and technology. Over 60 participants took part in Wilga 2024 meetings. Wilga Symposium embraces hardware and software technologies associated with photonics like optics, optical engineering, optoelectronics, electronics and electrical engineering, mechatronics, chemical and material engineering, applied physics, industrial solutions and applications. Around 50 papers were presented during Wilga 2024, oral and poster. Out of this number, some papers chosen by authors are published in this volume of Proc. SPIE mostly related to photonics. The major change was that during Wilga 2020, 2021, 2022 and 2024 Symposia, during summer editions, no traditional large young researcher sessions at large were possible to be organized. We hope that this will change and Wilga 2025 will return to a good tradition to organize concurrent social research sessions full of young researchers like was during Wilga 2019 and earlier.

The solution to TLC (Thin-Layer Chromatography) image segmentation in organochlorine pesticides (OCPs) toxicology research is crucial for scientific advancement and industrial applications. OCPs are found in minute concentrations in biological and environmental samples, necessitating precise segmentation for detection. Accurate segmentation is essential due to the potential multiple OCPs in samples, ensuring reliable component separation. Automated segmentation minimizes subjective errors inherent in manual methods, maintaining accuracy in toxicological studies. Standardized, automated processes ensure reproducible results, accelerating analysis for large-scale studies like environmental monitoring. Rapid TLC image segmentation reduces turnaround time critical for time-sensitive studies such as environmental assessments. Integration with analytical techniques like chromatography enhances comprehensive OCP analysis. Challenges include sample complexity with overlapping chemicals on TLC plates, and low OCP concentrations necessitating sensitive detection amidst noise and artifacts. Sophisticated algorithms are required for accurate analysis, sensitive to trace OCP signals and robust against experimental variations. Developing machine learning algorithms involves creating diverse training samples, a resource-intensive but essential task for reliable segmentation. This paper reviews segmentation methods, emphasizing threshold segmentation for its simplicity, speed, and accuracy. Algorithms for signal correction and peak detection enhance clarity, with trapezoidal integration accurately measuring peak areas. Developed in R, the program processes chromatographic images, providing tools for analysis, method refinement, and validation. These advancements are pivotal for ensuring the accuracy, efficiency, and reliability of OCP toxicology studies, crucial for environmental protection, public health, and regulatory compliance.

The theoretical principles of the Mueller matrix reconstruction of the parameters of optical anisotropy (linear and circular birefringence) of the myocardium polycrystalline myosin fibrillar networks architectonics are considered. The main signs of polarization manifestations of pathological transformations of the spatially structured myosin fibrils networks have been determined. The found tomography algorithms were used to process the results of myocardium native histological sections Mueller matrix mapping of those who died with long-term consequences of COVID 19. As part of the statistical analysis of the obtained native histological sections myosin fibrillar networks maps of linear and circular birefringence the most sensitive markers (asymmetry and kurtosis) to pathological changes in the myocardium. On this basis, differential diagnosis with excellent (>95%) balanced accuracy of long-term consequences of COVID 19 has been implemented. Further promising areas of research using the principles of laser-induced autofluorescence polarimetry of biological tissues are indicated.

A Mueller matrix model of biological tissues optical anisotropy is proposed. The relationships between the histological sections phase (birefringence) anisotropy parameters and polarization maps are determined. The experimental results of the brain and liver native histological sections polarization mapping of those who died with different durations of blunt mechanical trauma are presented. As part of the statistical analysis of the set of native samples digital microscopic images ellipticity coordinate distributions. The main time scenarios for changes in the biological tissues optically anisotropic architectonics polarization map’s topographic structure were determined. On this basis, a polarization method has been proposed for determining the duration of blunt mechanical trauma. The possibility of detecting, using statistical polarization markers, the duration of a mechanical injury for 18 hours with an accuracy of 45 to 60 minutes has been demonstrated. Prospects for further research using laser-induced autofluorescence polarimetry methods that are more sensitive at the molecular level are presented.

A short theory of the processes of formation of the laser speckle fields polarization structure of biological tissues with optically anisotropic architectonics is presented. A method of polarization interference mapping of speckle fields with layerby-layer phase scanning of holographically reconstructed distributions of complex amplitudes of biological preparations of various morphological structures and physiological states has been proposed and experimentally tested - biopsies of the thyroid gland and blood films of rats with different severity of sepsis. Layer-by-layer maps of polarization ellipticity of native histological sections of thyroid biopsies from donors and sick patients were studied. As part of the statistical analysis of phase coordinate distributions of polarization ellipticity, objective markers for the differential diagnosis of thyroid pathology were found - nodular goiter, autoimmune thyroiditis and papillary cancer. Very good (~90%) and excellent (≥90%) accuracy of differential diagnosis was achieved. Statistical markers for temporal monitoring of septic changes in polarization azimuth maps of polycrystalline blood films of rats were found. An example of further research using laser-induced autofluorescence polarimetry methods is presented.

This article illustrates the method and optical system of integrated photostimulation of the human nervous system, which forms the therapeutic effect of integrated electromagnetic fields through the eye system. In this work are consider the directions of application of optical stimulation of the structures of the human nervous system due to the formation of influence by integrated flows of electromagnetic fields with different radiation frequencies. A detailed description of the operation of the proposed integrated optical system provides opportunities for the creation of new hardware techniques.

The methods of layer-by-layer interference phase scanning with digital holographic reproduction of azimuth and ellipticity polarization maps for differential diagnosis of benign (nodular goiter - NG) and malignant (papillary thyroid - PT) thyroid tumours are presented. All studies in representative groups of histological sections M and AC were conducted using a circular laser beam. This ensured the azimuthal invariance and repeatability of measurements of polarization maps NG and PT. Integrated, and layer-by-layer polarization maps of NG and PT were received. Values of statistical moments of 1st - 4th order, which characterize NG and PT samples coordinate distributions of azimuth and ellipticity polarization have been determined and systematized. A set of statistical parameters - diagnostic markers most sensitive to changes in the polycrystalline structure NG and PT has been determined. The results of information analysis - determination of accuracy of vector-parametric and polarization-interference methods of differentiation of benign and malignant thyroid tumours are presented.

A Jones matrix model of biological fluids dehydrated films (facies) optical anisotropy is proposed. The relationships between the blood facies phase (birefringence) anisotropy parameters and the Jones matrix elements real and imaginary components are determined. The experimental testing results of the blood facies polarization matrix thesiography polycrystalline architectonics method with linear and circular birefringence coordinate distributions reproduction are presented. Integral blood facies phase anisotropy maps are presented. Scenarios changes of 1st - 4th orders central statistical moments, which characterize blood facies phase anisotropy maps are determined and physically analysed. The results of the Jones matrix thesiography method clinical testing is presented. Very good and balanced accuracy level has been achieved in the temporal differentiation of the long-term outcomes of patients who have undergone COVID-19. The prospects for further research using an integrated approach based on polarization analysis of laser-induced auto fluorescent images of supramolecular networks of blood facies are indicated.

The paper analyses the main areas of application of mathematical methods in medical diagnostics, formulates the principles of diagnostics based on fuzzy logic. The basic structure of the MIS medical information system for assessing effectiveness of the formation of pathological and keloid scars was developed and the main recommendations for its design were put forward, namely: the selection and purpose of the system; construction of an algorithm for solving the problems of evaluating biomedical information and forming diagnostic and prognostic conclusions. The results of the study made it possible to obtain a conclusion about the high reliability of the obtained results during evaluation of effectiveness of the formation of pathological and keloid scars. The practical value of the work is realized in the possibility of using an automated medical expert system to solve the problems of medical diagnosis based on fuzzy logic for assessing effectiveness of the formation of pathological and keloid scars.

The paper discusses the pyramid methods of direct and inverse Q-transformation. Examples of method implementation based on numerical data transformation are considered. The work is dedicated to the development of a method of pyramid transformation based on contour Q-transformation for encoding and processing images. The method of pyramid Q-transformation and an example of its application are detailed, and the results of its software model are analyzed.

An improved simulation model of photon propagation in scattering biological tissue has been developed based on the Monte Carlo method, which makes it possible to comprehensively consider all the main optical processes occurring during the propagation of optical radiation through the BT. The model differs from the known ones in that it allows one to consider the polarization state of the incident radiation and the change in its optical characteristics by tracking the electric field in each interaction event. Based on the developed statistical model, simulation modeling of the processes of photon propagation in the skin and its individual layers was carried out, which made it possible to establish the scattering and transmission characteristics for each of the skin layers with model parameters considering the angles of incidence of radiation in the visible and near-IR ranges, and the thickness of the skin layers. The simulation also made it possible to determine the relative intensity of backscattered photons depending on localization.

The authors simulated and accompanied by optical measurements various available design options for constructing output stages of laser medical equipment in light of possible implementations under certain limitations in the design, production and operation processes. The design option for a flexible optical fiber proposed by the authors is characterized by its manufacturability and structural simplicity, ensures effective, repeatable transformation of homogeneity in the beam and, most likely, allows for complex optimization. In the case of relevance of light spots in the form of elongated rectangles, it may be interesting to use a combination of individual parallel beams and lens assemblies. When removing the restriction of mandatory beam collimation, the use of channel waveguides as an aperture is of particular interest (both from the point of view of reducing optical radiation power losses and variability of the performance of output cascades of laser equipment).

Modern biomedical engineering is characterized by the rapid growth of data volumes that require processing and analysis to support clinical decision-making. Information technology plays a key role in ensuring the high-performance processing of these large datasets, contributing to the increased accuracy and speed of clinical diagnoses, as well as more effective subsequent patient treatment. This article aims to review current approaches and technologies used for biomedical data processing and to rethink the approach to using big data in decision support systems. Special attention is given to machine learning methods that enhance data analysis efficiency. The data processing approach proposed in this article allows for an 10-12% increase in the accuracy of spinal pathology classification, confirming its feasibility in medical practice.

Medical and biological images have different physical natures. A person-operator can perceive them in different ways. Therefore, an important task is the adaptation of the image processing process to a specific user, that is, to a narrow task that is solved by the consumer of information. It is often not enough to present an object to an observer with the help of an ideal display system, since the information required for image analysis in order to search and identify objects, the determination of various quantitative characteristics can be manifested only as a result of digital processing. Based on the above, the paper proposes an automated method for extracting the contours of ultrasound images using the example of studying the state of the hip joint. This method makes it possible to construct conventional lines on the image of the hip joint in an automated mode with minimal expenditure of resources and subsequently to calculate the values of certain angular indicators, and using known methods of studying the condition of the hip joint, to estimate its degree of dysplasia. By automating the process, the reliability of diagnostics can be increased.

For many years, the medical community has used experimental and clinical observation methods to study the course of various diseases. This approach has a high degree of validity, but it also has some limitations, such as high costs and difficulty in controlling all influencing factors and individual variability. In this regard, there is a need to create new methods for studying and predicting the course of diseases. One of these tools is a mathematical model. This study considers the possibility of mathematical modeling of the effectiveness of anti-tuberculosis treatment, based on the levels of biochemical markers of tuberculosis, namely Human-beta-defensin-1 (HBD-1), ferritin and interleukin-6 (IL-6). Based on a study of the relationships between the effectiveness of anti-tuberculosis treatment and the levels of Humanbeta-defensin-1, ferritin and interleukin-6, a mathematical model was created that allows predicting the effectiveness of anti-tuberculosis treatment based on determining biochemical markers at the beginning of treatment and after 60 days of anti-tuberculosis therapy with sensitivity and specificity of at least 88%.

Nasal breathing is essential for respiratory function, influenced by the complex anatomy of the nasal cavity, which conditions the inhaled air by heating, cleansing and humidifying it. Disorders such as nasal obstruction or nasal septal deformities can lead to sinusitis, decreased sense of smell and sleep apnea. Septoplasty corrects these problems, but requires postoperative splinting. This study focuses on the development of personalized intranasal splints using advanced CT image processing. The software analyzes DICOM images, identifies septal distortions and calculates geometric parameters to create personalized splints. These silicone splints stabilize the nasal septum, promote airflow and enhance recovery comfort, seamlessly integrating with 3D printing technology for precise replication. This innovative approach improves surgical outcomes, providing a better quality of life for the patient after septoplasty.

Modern technologies for recording human physiological parameters are developing very dynamically. The use of optical methods for measuring human physiological parameters, such as glucose, bilirubin, pulse, and blood saturation, have become key tools in non-invasive diagnostics and monitoring of patients' health. These technologies utilize the properties of light to obtain accurate and fast results, which can improve the efficiency of medical services. The article discusses the basic principles of optical biosensors, their application in measuring these physiological parameters and integration with telemedicine diagnostic systems. The prospects for the introduction of the Internet of Things (IoT) and artificial intelligence (AI) to create telemedicine systems that provide continuous real-time monitoring of patients and a rapid response to changes in their health status are also considered. We propose a system structure that allows remote monitoring of key health indicators of patients, especially infants and children, as well as patients with chronic diseases such as diabetes and cardiovascular diseases. The introduction of such technologies helps to increase the availability of medical services, reduce the number of hospitalizations, and improve the quality of life of patients.

Since the advent of deep learning a decade ago convolutional neural networks have been the predominant method for approaching computer vision tasks. However, Transformer model, which has shown significant achievements in the field of natural language processing, is increasingly being applied to computer vision tasks and is demonstrating comparable or superior performance. The article discusses the application of Transformer model to the super-resolution task. The direct application of the original Transformer achieved performance comparable to the contemporary convolutional neural networks. However, the self-attention mechanism, which underpins Transformer model, involves quadratic computational complexity with respect to the size of the input image, presenting a significant challenge for processing high-resolution images. Further research has significantly improved performance, but these improvements are not exhaustive. An overview and comparative analysis of these studies are presented.

This article addresses the development of an intelligent military aircraft identification system using artificial intelligence, machine learning, and deep self-learning technologies to enhance national security and military efficiency. The system aims to automatically and accurately recognize and classify aircraft in images, offering advantages over traditional methods such as higher productivity, speed, accuracy, and the elimination of human error. The importance of deep learning solutions for threat detection and operational efficiency is emphasized. Modern visual data-based object recognition methods and tools are analysed. The methodology includes collecting and preprocessing data, developing a high-precision recognition system based on Yolov8, annotating objects with Roboflow, and creating training, validation, and testing subsets in the yolo format. The paper details the dataset formation process and presents satisfactory results in fast recognition of military aircraft with high classification accuracy. A comparative analysis of Yolov8, R-CNN, and GPT-4 models shows Yolov8's superiority in prediction accuracy and performance. The article describes the model management system for adjusting hyperparameters, selecting object categories, and initiating the training and forecasting process. Testing results demonstrate Yolov8's optimality for military aircraft identification, achieving accurate target identification in complex situations using advanced deep learning algorithms.

This article illustrates the problem ensuring homogeneity and uniformity of nanoscale coatings obtained in vacuum by carbon vapor deposition. Carbon vapors are obtained by heating in a vacuum with one or more electron beams from a water-cooled crucible, which is covered with a molten layer of refractory metal. Due to the peculiarities of the trajectory of the movement of carbon vapors from the crucible to the substrate, a situation may arise with the formation of a "visor" type defect on the coating, which prevents the uniformity of the coating.

The problem of inverse correlation filters design to recognize a set of objects is considered as the problem of regression parameters estimation on the base of input data arrays and desirable response. The data and response should be processes with zero mean to consider this problem as evaluation of regression parameters. The problem is solved using the least squares method with regularization. The regularization is optimized to achieve high resolution of the filters in conjunction with capture’ broad band of objects given by a set of templates. The least squares method is using in the terms of singular value decomposition that made it possible to linearize the nonlinear ridge regression optimization problem. The methods to false recognitions elimination are considered, It was shown that the regression approach gives additional condition to recognize classes of objects. This allows to have more high accuracy in recognition of desired objects on a foreign background in comparison with other correlation filters types.

This article presents the implementation of a modified parallel-pyramidal algorithm for the efficient processing and identification of images. The method involves creating a systemic model that supports spatial, temporal, and network data integration, forming a dynamic pyramidal-hierarchical network. The article details the techniques for sorting vectors, G-transformations for modifying vector elements, and the shifting procedure that facilitates efficient data transformation. The described procedures are integrated into the overall data processing sequence, which includes the iterative application of these methods until the final result is achieved.

The paper reveals a method of highly productive determination of normalized vectors for the surfaces of threedimensional objects. The method is based on the approximated calculation of even vectors of the rasterization line by adding the odd neighboring unit vectors. For the determination of further need in the normalization of obtained vectors, the computation of special threshold metrics is proposed. For the accelerated calculation of the threshold metrics, the developed expressions are given. In case of normalization of even vectors, it is recommended to use the developed polynomial approximate expressions. The plots of relative errors between obtained simplified and reference expressions are given. The possibility of increasing the productivity of the method by calculating the shared vector coordinate increments for each rasterization line is analyzed. The experimental research of productivity gain from the new method usage is carried out. During the study, the six variants of possible method usage are considered. The results of the research analysis are given in the table. The new method is designed for usage in highly effective visualization systems.

This work is devoted to the development of a new image classification method based on the application of a hybrid semi-supervised learning algorithm for convolutional neural networks. The proposed method is based on the active learning approach and uses metrics to identify and label samples that have a high impact on the certainty metric of the entire neighborhood. To evaluate the effectiveness of the proposed method, a synthetic semi-supervised benchmark dataset is used. The optimal choice of the graph algorithm and clustering algorithm is carried out by testing different combinations of algorithms on the synthetic dataset. The proposed method demonstrates an increase in accuracy compared to the base classifier and more efficient utilization of active labeling compared to supervised active learning. As a result, the possibility of identifying and solving the unknown class problem is realized.

In this study, the emissivity values of metal powder were examined by measuring reference temperature values using thermocouples, an infrared camera and aluminium foil to determine the reflection temperature. This enabled the testing of a methodology for determining emissivity in order to implement an intelligent control approach in additive manufacturing. The research established emissivity values for the surface of 316L powder steel that range from 0.33 to 0.46 in the temperature range from 50 to 600°C. The proposed approach allows the calibration of an infrared camera to accurately determine the temperature values of metal surfaces, which opens up the possibility of using the measurement results for intelligent control of laser power in additive manufacturing.

This article illustrates the application of the Kalman filter in real-time human re-identification tasks to improve accuracy, reliability, and reduce computational costs while determining a person's position and orientation. The use of the Kalman filter addresses noise filtering and prediction issues in human re-identification tasks.

Matrix multiplication is a fundamental operation in various scientific and engineering applications, often demanding high computational resources. With the growing complexity of data-intensive tasks, optimizing this operation to run efficiently on hardware accelerators like FPGAs has become crucial. However, the challenge lies in balancing the resource constraints while maximizing performance. In this project, we address this problem by developing High-Level Synthesis (HLS) software tailored for efficient matrix multiplication on FPGAs. Our solution focuses on resource optimization, adhering strictly to predefined constraints to ensure optimal performance. By leveraging the parallel processing capabilities of FPGAs, our implementation significantly reduces computation time compared to traditional methods. We present the design and implementation of our HLS-based matrix multiplication solution, highlighting key strategies for resource management and optimization. The results demonstrate a substantial improvement in performance, showcasing the effectiveness of our approach in real-world applications. This article summarizes our methodology, experimental setup, and the performance gains achieved through our optimized solution. This article presents the results of matrix multiplication on FPGA systems with Resource Optimization and Constraints.

The article improves the efficiency of counteracting unauthorised access by using algorithms and a protocol for monitoring information security in fibre-optic transmission systems and access networks with the help of the proposed software, which functions as part of a combined (hardware and software) method of protecting optical information flows in FOCLs. The principles of building network protocols are investigated. The functionality of the developed algorithms and the protocol for monitoring the information security of fibre-optic transmission systems is substantiated. The algorithms for stopping and restoring the transmission mode for the subscriber and linear sections of the optical network based on GPON and GEPON technologies, as well as the protocol for monitoring the information security of fibre-optic transmission systems are developed. The practical significance of the work is to improve information security through the use of a combined security control tool that operates on the basis of the developed algorithms and the information security control protocol.

The article is dedicated to advancing technologies in chaotic optical communication and investigating chaotic laser generation in measurement systems and instruments. It conducts an analysis of the informational potential of chaotic signals in the optical range and explores the feasibility of implementing chaotic dynamics in lasers. It is demonstrated that among the current challenges are precision and stable generation of chaotic laser modes, as well as precision and stable synchronization of chaotic laser generators. The definition of chaotic modes is proposed as a set of characteristic types of chaotic oscillations that are stable, distinguishable, and reproducible within an oscillatory system, along with a method for estimating the number of modes and their designation. A method for modulating pump energy using a Chua generator is suggested. Additionally, a model for precision control of chaotic laser generation is proposed, allowing for the generation of chaotic laser modes with the ability to measure and control laser emission parameters and chaotic dynamics based on monitoring small changes in control parameters and influencing factors.

Currently, more and more attention is being paid to monitoring the state of road infrastructure, since it is the public domain of the country. As a result of the large flow of traffic and environmental influences, it leads to wear of the road surface. Timely detection of defects and damage to the pavement allows us to develop effective programs for maintenance and restoration of the pavement, which allows us to achieve significant costs and save time. Currently, fiber sensors are used as a more effective tool for monitoring the condition of the road surface. Since road surfaces are multilayer structures consisting of layers of granular and bituminous materials, the sensors must correspond to the heterogeneous nature and mechanical properties of the pavement materials. Fiber sensors based on the fiber Bragg grating (FBG) meet these requirements. Fiber sensors based on FBG are widely used due to advantages such as long-term stability and durability, good geometric shape, versatility, corrosion resistance, electromagnetic interference protection, low cost and high detection accuracy.

The creation of modern intelligent systems based on artificial intelligence requires solving the complex problem of highspeed processing of large amounts of information. The article substantiates the direction of perspective development of optoelectronic means combining electronic processing with optical input/output of data arrays. An example of the implementation of an optical channel classifier based on smart-pixel matrices is given as part of an optoelectronic classifier. Each smart-pixel matrix has electrical control inputs and the ability to process the input optical signals electronically with the output of the result in the form of optical signals.

This article discusses the methodology for creating digital content for educational programs in photonics and optoelectronics. The primary focus is on using the SCORM 2004 standard, which enables the development of high-quality, standardized, and interactive educational materials. The work outlines the content development stages, including studying the standard, planning, creating educational resources, integration, testing, assembly and publication, monitoring, and updating. The effectiveness of the methodology was evaluated using the COR (Completion Rate) and NPS (Net Promoter Score) metrics. An example implementation of the methodology is presented in the "Computer Science" course developed for students in the photonics and optoelectronics educational program. The results showed a high level of course popularity and demand, as evidenced by a COR indicator of 68%. The article concludes that the proposed methodology is viable and effective, with plans for its expansion to other educational programs.

This paper presents an implementation of a modified parallel-pyramidal algorithm for efficient image processing and identification. The method involves the creation of a system model that supports the integration of spatial, temporal and network data to form a dynamic pyramidal-hierarchical network. The paper details vector sorting techniques, Gtransformations for modifying vector elements, and a shifting procedure that facilitates efficient data transformation. The procedures described are integrated into a general data processing sequence that involves iterative application of these methods until the final result is achieved. How the algorithm works is shown in the example of laser beam projection analysis.

Wireless networks are crucial to modern communication infrastructure, supporting a wide range of applications from personal use to industrial operations. The rapid growth of mobile devices, IoT, and advanced technologies like 5G has heightened the demand for scalable, efficient, and reliable wireless networks. Routing in wireless networks is particularly complex due to the dynamic nature of the medium, device mobility, and diverse topologies. Traditional routing algorithms, though effective in some cases, frequently fail to handle the complexities of contemporary wireless networks. Integrating machine learning technologies with routing algorithms offers a promising solution. In particular, reinforcement learning (RL) has become prevalent in routing due to its ability to operate without the need for extensive datasets. This article evaluates two RL-based routing algorithms, by comparing their average delivery times under various loads across various topologies. This study offers a comprehensive analysis that addresses the importance of different network structures, providing insights into how each algorithm performs across diverse topologies, an aspect not covered in previous surveys.