Solution of the problem of recognition and vectorization of parts of sea transport requires formation of skeletonized images, homotopic (geometrical primitives, topologically equivalent in shape and their coherence) to parts' shapes. The author has performed a comparative analysis of the best methods of parallel, topological skeletonization of the area objects, based upon application of space extractors. The analysis showed that the methods existing in the investigated objects zone possessed typical drawbacks, expressed in iterative distortions of primitive topology and their compositions. The objective of the article is to through the light upon the developed methods of improvement of topological equivalence of the resulting skeletons to the shapes of the parts of sea transport, by means of gradual correction of typical distortions of skeletons. The developed methods assumes correction of skeleton's iterative distortions by modified extractors of the principal method of skeletonization and restoration of the resulting skeleton by extractors of restoration of homotopic skeleton, on the basis of developed rules of its reconstruction. Execution of the proposed method was carried out on example of the basic method Wu R.Y. & Tsai W.H. Examples of the results of skeletonization of parts' drawings were given, verifying efficiency of the proposed methods. The methods can be adapted to the methods of topological skeletonization of area objects, based upon application of space extractors.

УДК 004.93Молчанова В. С. Ст. преподаватель кафедры информатики ГВУЗ «Приазовский государственный технический университет»,Мариуполь, Украина АДАПТИВНЫЙ ПОРОГОВЫЙ МЕТОД БИНАРИЗАЦИИ РАСТРОВЫХ ИЗОБРАЖЕНИЙ ТЕХНИЧЕСКИХ ЧЕРТЕЖЕЙВ статье предлагается метод адаптивной пороговой бинаризации растровых изображений технических чертежей. Проанализированы специфические особенности изображений технических чертежей и их влияние на результат бинаризации известными универсальными методами. Выявлены артефакты, возникающие на результирующих бинарных изображениях и предложены способы их устранения.В основу предложенного метода положено соображение о том, что очень маленькие фрагменты изображения чертежа напоминают букву «С», которая в особых случаях вырождается в прямую линию. Дальнейшие рассуждения сводятся к выбору способа определения соответствия объекта на изображении форме «С». Отдельное влияние уделено вопросам настройки порога бинаризации в зависимости от яркости изображения.В работе представлена математическая модель предложенного метода, его алгоритмическое описание. Проведены эксперименты по исследованию качественных и количественных показателей эффективности. Качественная оценка выполнялась путем сравнения результата бинаризации с эталонным черно-белым изображением. В качестве критериев количественного оценивания рассмотрены такие критерии, как время выполнения алгоритма, полнота и f-мера. Представлены результаты проведенных экспериментов, показывающие превосходство предложенной в работе методики бинаризации растровых изображений технических чертежей, как количественно, так и качественно.Ключевые слова: растровое изображение, бинаризация, технический чертеж, порог бинаризации, яркость, оттенок серого, полезный сигнал, фон, шум, ошибка

is used in image processing of technical drawings, including drawings of sea transport parts, since the object's skeleton reflects its topological structure. Сomparative analysis of the best methods of parallel topological skeletonization of the area objects, using spatial masks, showed that they give iterative distortions to the topology of primitives and their compositions. Therefore, the task of developing a technique for homotopic skeletonization of bit-mapped drawings of sea transport parts is relevant. Objective. To develope technique of improvement of topological equivalence of the skeletons to the сontour of sea transport parts, by means of gradual correction of typical skeleton's distortions. Method. Сorrection of skeleton's iterative distortions by modified spatial masks of the basic method of skeletonization and the reconstruction of the resulting skeleton by masks to restore its homotopy to the original, on the basis of developed reconstruction rules. Execution of the proposed technique was carried out on example of the basic method R.Y. Wu & W.H. Tsai. Results. The proposed technique is implemented as a program application that allows to perform quality skeletonization of images of drawings of sea transport parts. Conclusions. The shown examples of results of skeletonization of drawings of parts confirm efficiency of the proposed technique. The technique can be adapted to the methods of topological skeletonization of area objects, based upon application of spatial masks.

No abstract

Context. Typically, interaction between user and mobile devices is realized by touchings. However, many situations, when to implement such interaction is too awkward or impossible, exist. For example, with some diseases of musculoskeletal system, motility of movements may be impaired. It leads to inability to use device efficiently. In that case, a task of looking for alternative ways of person-device interaction becomes relevant. Voice interface development can be one of the most prospective tasks in that way. Objective. The goal of the study is to develop a project of neural network architecture and internal components for voicecontrolled systems. Resulting interface have to be adapted for processing and recognition Ukrainian speech. Method. An approach, based on audio signal analyzing by sound wave shape and spectrogram, is used for making got via microphone data, appropriable for processing. Using neural network makes possible sounds classification by generated audio signal and information of its transcription. The neural network structure is completely adapted to peculiarities of Ukrainian phonetics. It takes into account the nature of the sound wave, generated during sound pronunciation, as well the number of sounds in Ukrainian phonetics. Results. Experiments were carried out aimed to choosing optimal neural network architecture and training sample dimension. The root-mean-square deviation of neural network error was used as the main criterion in assessing its effectiveness. A comparative analysis of effectiveness of the proposed neural network and existed on the market speech recognition tools showed improvement in the relative measures of recognition by 9.26%. Conclusions. Obtained in the research results can be used for full-featured voice interface implementation. Despite the fact that the work is focused on recognition Ukrainian speech, the proposed ideas can be used during developing transcribing services for other languages.