Defense of the dissertation of Kerimhan Bekjan Temirhanuly for the degree of Doctor of Philosophy (PhD) in the specialty «6D070400 - Computer engineering and software»

L.N. Gumilyov Eurasian National University, a dissertation defense for the degree of Doctor of Philosophy (PhD) by Kerimhan Bekjan Temirhanuly on the topic Development of software for video sequence analysis based on descriptive-associative algorithms by specialty «6D070400 - Computer engineering and software».

The dissertation was carried out at the «Computer and Software Engineering» of L.N. Gumilyov Eurasian National University.

The language of defense is kazakh

Official reviewers:

Mamyrbaev Orken Zhumazhanovich - doctor of philosophy (PhD), specialty: 6D070300 - Information systems, associate professor, deputy general director of the "Institute of Information and computational technologies" of the Science Committee of MSHE RK (Almaty, Republic of Kazakhstan);

Buribaev Zholdas Alladinovich - doctor of philosophy (PhD), specialty: 6D075100- Informatics, Computer science and management, associate professor of the Department of computer science, al-Farabi Kazakh national university (Almaty, Republic of Kazakhstan).

Temporary members of the Dissertation Committee:

Shashi Bhushan - doctor of philosophy (PhD), specialty: Computer science & engineering, Universiti Teknologi Petronas (Seri Iskandar, Malaysia);

Yerimbetova Aigerim Sembekovna - doctor of philosophy (PhD), candidate of technical sciences, specialty: 6D070300 - Information systems, 05.13.17 - Theoretical foundations of computer science; Leading Researcher of the “Institute of Information and computational technologies” of the Science Committee of MSHE RK, (Almaty, Republic of Kazakhstan) (instead of Pak Alexander Aleksandrovich - candidate of technical sciences, specialty: 05.13.18 - Mathematical modeling, numerical methods and software packages, associate professor, head of the laboratory “Intellectual analysis of big data” of the Institute of information and computational technologies of the Science Committee of MSHE RK (Almaty, Republic of Kazakhstan) in accordance with the decision of the Dissertation Council of L.N. Gumilyov Eurasian national university in the direction of personnel training 8D061 - Information and communication technologies, educational programs: "6D070400 - Computer Science and Software", "8D06104 - Computer Science and Software", Protocol №4 dated 4.12.2024.);

Kumargazhanova Saule Kumargazhanovna - candidate of technical sciences, specialty: 05.25.00 - Information systems and processes, associate professor of the School of digital technologies and artificial intelligence of the D. Serikbaev East Kazakhstan technical university (Ust-Kamenogorsk, Republic of Kazakhstan).

Scientific advisors:

Zhumadillaeva Ainur Kanadilovna - candidate of technical sciences, associate professor of the Department of Computer and software engineering of the L.N. Gumilyov Eurasian national university (Astana, Republic of Kazakhstan);

Nedzved Alexander Mikhailovich - doctor of technical sciences, professor, head of the Department of Management information systems, faculty of applied mathematics and informatics, Belarusian (Minsk, Belarus).

The defense will take place on December 23, 2024, at 11:00 in the Dissertation Council for the training direction «8D061 - Information and communication technologies» in the specialty «6D070400 - Computer engineering and software» of L.N.Gumilyov Eurasian National University. The Dissertation council meeting will be held offline and online on the Microsoft Teams.

Link to Microsoft Teams: https://clck.ru/3EaNXG 

Address: 010000 Kazakhstan, Astana, A. Pushkin str., 11, educational building, auditorium 222.

Abstract (English):ANNOTATION of the PhD Thesis on the Specialty «6D070400 - Computer Science and Software Engineering» Bekzhan Temirhanuly Kerimkhan «Development of software for video sequence analysis based on descriptive-associative algorithms» Relevance of the research topic. Nowadays, image is actively used as a means of presenting the results of a wide variety of studies in many fields. Visual interpretation of microscopic preparations is central in medical diagnostics, and interpretation of space images is actively used in many tasks of cadastral agency, natural resources monitoring and other tasks. The human visual system is good at evaluating qualitative characteristics of objects, but quantitative description of the same objects in most cases is rather subjective. Automation of counting of objects and measurements of their characteristics allows not only to increase the accuracy of object assessment, but also allows to store images and results of their processing in a database of large capacity, hence, to use large amounts of data in diagnosis, which, as well as unification of measurements, allows to make the diagnosis quite objective. Medical and satellite images contain many complex structures characterizing the properties of the studied objects. At the same time, there are no effective methods of their analysis that take into account such specificity of images. Traditional technologies are mainly focused on image enhancement methods and do not take into account the features of objects. The modern direction of image analysis based on neural networks requires preparation of quality sets and is limited by local space convolution operations. Therefore, the development of new methods and software tools for analyzing objects in images is an urgent scientific and practical task. Existing technologies for analyzing complexly structured images represent a set of disparate sets of operations. They are mainly designed for image enhancement and private tasks that require simultaneous participation of at least two specialists. One defines the specifics of the subject area, the other defines the set and sequence of image analysis operations for the study. Such technologies are inefficient in complex automation. The software market has developed a large number of product for solving a narrow class of tasks with its own set of concepts, definitions and unique user interface. The use of such a product is quite complex. Nowadays, automatic image processing is one of the most important directions in the field of artificial intelligence performing pattern recognition. One of the most effective tools for pattern recognition is systems built on descriptive-associative algorithms. In the system of technical vision (SVS) it is necessary several methods and algorithms that solve the same problem in different ways, while providing the necessary performance in terms of speed and reliability of identification. For example, such tasks include analyzing blood flow, which reflects eye health and is impaired in many diseases. Many pathological processes occur at the cellular level, such as blood microcirculation in blood vessels, and medical image processing is a challenging recognition task. Blood vessels in retinal images contain significant information about pathologic changes caused by ophthalmic diseases such as diabetes, hypertension, and atherosclerosis. Computerized retinal image analysis plays an important role in the diagnosis, treatment, screening, evaluation, and clinical study of ocular diseases. However, automatic segmentation of retinal anatomical structures is challenging due to the presence of lesions and noise, irregular illumination, intensity drift, missing image contrast, variable vessel width and central vessel reflex. A considerable amount of work has been done on automated retinal vessel segmentation, which can be categorized into methods based on matched filtering, morphological processing, vessel tracking, multiscale analysis, pattern recognition and model-based algorithms. Existing methods of blood flow measurement are limited due to complex assumptions, equipment requirements and calculations. In this paper we propose a methodology for descriptive analysis of video sequences, where the motion is decomposed into parts by key features, this methodology is tested on the task of determining the characteristics of blood flow in the vessels of the conjunctiva of the eye, such as linear and volumetric blood flow velocity and topological characteristics of the vascular network. The methodology is performed in stages. In the first stage, whole video stream processing is performed to improve the analysis conditions. The cascade of sequential processing of fragments is started to refine motion features on the basis of integral optical flow to determine dynamic characteristics of motion at different levels (background, group of objects, object, motion inside the object). In the practical task of blood vessel motion, these features allowed us to identify fragments of blood flow velocity changes in the blood vessels of the eye. We show the effectiveness of our method in scenes with natural blood vessels of the eye. The study provides valuable information to novices with limited experience in diagnosis and can serve as a valuable tool for experienced medical professionals. Dissertation Research Objective. The purpose of the research is to develop and experimentally investigate the methodology of image sequence analysis by means of motion parsing on the basis of descriptive algorithmic schemes, intended for the technology of software development or operational correction in case of complex motion in video sequences containing many dynamic objects. Objectives of the study: 1. To analyze scientific research in the field of dynamic objects. 2. To determine the properties of the dynamic object and descriptors of its motion as a subject of monitoring. 3. To develop a methodology for monitoring dynamic objects on the basis of motion descriptors 4. Develop a methodology for monitoring a dynamic scene. 5. To realize and test the technology of blood flow analysis monitoring based on the technique of descriptive-associative analysis of dynamic scene. The main object of the research is descriptive-oriented methods for extracting information and motion data from video sequences to build a unified video monitoring software architecture. Scientific novelty. The developed methods are new and original, they have no foreign analogues and, with respect to video sequence analysis, currently determine the world level. They differ from the known ones in that they use a descriptive approach to analyze motion on video sequences, which allows us to disassemble the motion into parts, for which to determine groups of dynamic objects. This allowed the characterization of individual motion for them. A significant distinguishing feature is also the formalization of motion representation as an integral step in both analysis and evaluation to form adaptation sets of self-modifying software commands. The main provisions put forward for defense: 1. Formalization of a dynamic object on the basis of setting descriptors of its motion, which allowed to use in the definition of motion formulations for grouping into groups and sets, allowing to simplify definitions and solutions of problems of monitoring objects on video sequences. 2. Methodology for monitoring dynamic objects on the basis of motion descriptors and sequential analysis of complex-structured multi-temporal images, based on the formation of descriptor maps, which allow to determine the signs of object behavior, necessary for further decision-making in monitoring systems 3. The technique of descriptive monitoring of a dynamic scene, based on the division of dynamic objects into groups and the use of descriptive motion maps to compensate for the motion of the background, unimportant for the analysis of objects and the motion of dynamic objects proper, for which both external and internal motion is taken into account. 4. The technology of blood flow analysis based on the technique of descriptive analysis of the dynamic scene of sclera as an example of a practical problem using the technique of descriptive monitoring in which the motion of background, vessels and blood cells is present. As a result, blood flow is defined as the main dynamic object whose motion is specified by combining the dynamic descriptors of a group of cells. The definition of dynamic descriptors of vessels and background allowed to compensate all external motion and focus exclusively on the study of blood flow properties. Study Subject. Algorithms and methods for analyzing complex-structured motion of objects in a dynamic scene. Research Methods. Algorithms and methods of computer vision and image comparison Practical significance of the obtained results. Practical significance of the obtained results - Descriptive description of the motion analysis technique is designed to perform motion analysis and can be used in automatic object monitoring systems. Recommendations for the practical use of the results are to use the results of the study for monitoring and quantitative analysis of changes occurring in a given area or scene. The formalization of the task of motion observation on video sequence of images is proposed, which will significantly simplify the process of forming a scheme for solving the tasks of monitoring dynamic objects by video sequence. Recommendations for further development of the research: it is planned to continue the research in terms of development of algorithms for motion analysis in different research conditions, as well as acceleration of already developed algorithms. The developed algorithms can find wide application in the field of monitoring of moving objects not only in medicine, but also for controlling the behavior of crowds or traffic flows. Approbation of the results of the dissertation. The main results of the dissertation work were reported at seminars of the L.N. Gumilev Eurasian National University, at international conferences held in Russia, Belarus: 1. Seminars of doctoral students of the Department of "Computer and Software Engineering", Astana, 2019-2021. 2. International Scientific Conference "Global science and innovations 2019: Central Asia", Nur-Sultan, 2019. 3. International Scientific Conference "Ǵylym jáne Bilim - 2022", Nur- Sultan, 2022. Publications published based on the results of the study, including: - In scientific journals indexed in the Web of Science and Scopus database: 1. B. Kerimkhan, A. Nedzved, A. Zhumadillayeva, K. Dyussekeyev, G. Uskenbayeva, B. Sultanova, L. Rzayeva. Automation of flow analysis in scleral vessels based on descriptive-associative algorithms. // Scientific Reports, -2023. -13(4650). - Articles published in scientific journals recommended by Committee for Quality Assurance in the Sphere of Education and Science of the MES RK: 1. Kerimkhan B., Nedzved A., Zhumadillayeva A., Duisenova G. Методика автоматизации анализа больших наборов изображений для задач мониторинга // Bulletin of the National Academy of Sciences of the Republic of Kazakhstan. - Almaty 2022. - №2 (84). - 82-89. 2. Kerimkhan B, Zhumadillayeva A., Nedzvedz A. Analysis of dynamic changes from large set of remote sensing images // Scientific Journal of Astana IT University. - Astana 2022. - №11. 4-13. 3. Kerimkhan B., Zhumadillayeva A., Nedzved A., Ilyasova M. Динамикалық обьектілердің қозғалыс карталары және орындалатын әрекеттер // Bulletin of the University of Toraigyrov. - Pavlodar 2022. - No.4. 152-164. - In the proceedings of international conferences: 1. Kerimkhan B.T. Дескрипторларды құру және суреттерді салыстыру. // Collection of materials of the VI International scientific and practical Conference "GLOBAL SCIENCE AND INNOVATIONS 2019: CENTRAL ASIA", Nur-Sultan, 2019., 225-228s. 2. Kerimkhan B.T. Бейнелерді салыстыру және динамикалық обьектілерді анықтау. // Student men zhas galymdardyn "ǴYlym jáne Bilim - 2022" Collection of reports of the XVII International Scientific Conference, Nur-Sultan, 2022., 943-948c. - Certificate of state registration for the copyright object. Computer program: Descriptive-associative algorithmder negizinde scleraldy tamyrlardagi agyndy taldaudy automattandyru. Certificate of State registration of copyright rights, No. 24788 dated April 5, 2022. - The act of implementing the developed software product: the software "Descriptivtik-associative algorithmder negizinde scleraldy tamyrlardagi agyndy taldaudy automattandyru" (authors Kerimkhan B.T. and Zhumadillaeva A.K.) is currently used by the medical clinic to assess the state of the microcirculatory bed of other organs and systems (03/16/2023). The scope and structure of the dissertation. The dissertation work consists of an introduction, three chapters, conclusion, 3 appendices. The work is executed in printed form on 97 pages, with application of computer capabilities of emphasis in the form of illustrations, schemes and tables. The list of references consists of 103 titles. The author expresses deep gratitude to the scientific supervisor, Associate Professor of the Department of "Computer and Software Engineering", Ph.D. Zhumadillaeva Ainur Kanadilovna and foreign consultant, Professor, Ph.D. of the Belarusian State University (Belarus, Minsk) Nedzved Alexander Mikhailovich for invaluable work and consultations during the research.

Dissertation

List of Scientific Works

Consultant's Review

Foreign Consultant's Review

Conclusion of the Research Ethics Committee

Official Reviewer's Review

Official Reviewer's Review