Currently there is a contradiction between availability of various new equipment, which provides a stream of digital video data, in particular in the form of point clouds from mobile laser scanning, and the lack of adequate efficient methods of information extraction and analysis. This project is aimed at resolving this contradiction on the basis of neural modeling field theory and dynamic logic (DL) proposed by L. I. Perlovsky. The main result of the project will be a method of extracting information from digital video data in the form of hybrid clouds of mobile laser scanning data points for their analysis based on neural modeling field theory and DL. The success of this project depends on the successful integration of approaches from various fields of science and technology (interdisciplinarity): artificial intelligence, pattern and object recognition, logic, algorithm theory. The significance of the development of the proposed method is to create a fundamental theoretical basis for new application algorithms and software in the field of autonomous driving, "smart city" projects, ensuring safety for sites of various purposes, etc. The scientific novelty of the proposed method is that it will solve, by a fairly new method, the relevant problem of extracting and analyzing information from a not particularly traditional type of digital video data represented by a hybrid cloud of laser scanning points. This will allow to significantly expand the existing boundaries of knowledge in the field of extraction and analysis of information from various digital video data. The main hypothesis of the research is that the new method based on L. I. Perlovsky's neural modeling field theory and DL will improve the performance of relevant calculations and close the existing gaps in the use of various digital video data.
Short CommunicationDigital image and video processing are one of the most resource-intensive tasks and at the same time -important and relevant, closely related to the problem of artificial intelligence (AI) and modeling of cognition. Its frequent combinatorial complexity complicates the already difficult task: partition and recognize -partition video data (video image) into meaningful objects and recognize their essence, classify them. Mathematical logic offers to solve recognition problems by iterating over all known images and comparing them with the presented one. Even with a modestly sized set of options, the number of combinations to compare would be huge. At the same time, there are many separate methods and concepts in this field, often used together: decision trees, support vector method, convolutional neural networks, deep learning, K-means, latent Dirichlet allocation, etc. L. I. Perlovsky has developed an original approach to cognition modeling based on the theory of neural modeling fields (NMF) and dynamic logic (DL) [1-3] that we propose for analyzing various digital video data. While ordinary logic operates with precise statements like "this is a chair", DL is a vague-to-crisp process.DL implements...