Search academic texts

Information × Registration Number 0212U006561, 0112U000921 , R & D reports Title Development of tools for creation of intellectual platforms for recognition of visual-speech images with elements of semantic perception and content-analysis of information texts. popup.stage_title Head Shevchenko Anatoliy Ivanovich, Registration Date 14-05-2012 Organization Institute of Artificial Intelligence Problems popup.description2 The object of research is a method of data sequences storing, finding and processing by content. The subject of research - methods and tools for automated processing of symbolic and image information using the method of data sequences storing, finding and processing by content. Research goal - development of tools for creation of intellectual platforms for recognition of visual-speech images. Research methods - pattern recognition, computational linguistics, methods of data storing and finding, computer simulation. Theoretical results. There are found the parameters of the method, which allows to reduce the cost of memory, as well as to reduce the time required to perform basic operations. It is pointed out prospects of the method for solving speech recognition task and other applications of artificial intelligence. A new approach to finding of speech signal boundaries proposed. Methods for speech recognition by allocation of characteristic features in the speech signal have been further developed. Recognition algorithms adapted to detect the sounds of the Kazakh vowels. Approach to pattern recognition based on interphoneme transitions finding and recognirion proposed. The technique of comparative analysis of speech recognition created. Study of DTW-modifications in task of isolated words recognition carried out. A list of vocabularies and databases needed to automate the semantic and syntactic analysis of texts was identified. An algorithm for parsing simple common uncomplicated sentences and the general scheme of semantic and syntactic analysis of arbitrary sentences were proposed. The practical results. A library of morphological analysis without dictionary for Russian words and an experimental system of morphological marking for Russian texts were created. These systems, as well as the library of declarative morphological analysis of Russian words RDMA_IAI were tested. The vocabulary database of library RDMA_IAI was improved. The effectiveness of implementation. The method of comparative analysis of speech recognition methods used while working under a contract for the development of scientific and technical products. The library RDMA_IPII is used while working under contract for services. Field of of use. The development of intelligent systems to analyse natural language text and speech recognition. Product Description popup.authors Єрмоленко Тетяна Володимирівна Гнитько Дмитро Сергійовіч Дорохіна Галина Володимирівна Капліна Олеся Олександрівна Ніценко Артем Володимирович Шевченко Анатолій Іванович Шелєпов Владислав Юрійович popup.nrat_date 2020-04-02 Close