CC BY
15
УДК 62-503.55
ОБРАБОТКА МЕДИЦИНСКИХ ИЗОБРАЖЕНИЙ
А. Ю. Тычков, П. П. Чураков, А. В. Кузьмин
PROCESSING MEDICAL IMAGES A. Y. Tychkov, P. P. Churakov, A. V. Kuzmin
Аннотация. В статье показано, что для эффективной обработки изображений необходимы методы, способные точно выделить объекты интереса без явных искажений его деталей и границ. Предложено для точного выделения объектов интереса на малоконтрастных изображениях использовать метод декомпозиции на эмпирические моды. Разработан алгоритм выделения контура сердца на флюорографических изображениях, основанный на текстурной сегментации.
Ключевые слова: медицинское изображение, цифровая обработка, декомпозиция на эмпирические моды.
Abstract. In article is shown that for efficient processing of images necessary methods, capable exactly to select the sidebar without evident distortion of its details and borders. It is offered for edge extraction of the sidebar on fluoro to use the method to decompositions on empirical modes. The designed algorithm edge extraction of the sidebar heart, founded on texture of the segmenting.
Key words: medical image, digital processing, empirical mode decomposition.
Medical images obtained by means of digital x-ray photographic systems is today among the principal sources of diagnostic information on the physiological slate of the thoracic organs in man and animals. The Digidclca x-ray device from the firm of Oldclft (Holland). The Pulmoskan low-dosage digital photofluorograph from the firm of Adani (Belarus), and the Sibir-N low-dosage digital x-ray device (Russia) is the most well-known of these digital x-ray photographic systems [1, 2].
Modem digital x-ray photographic systems must assure high image quality and extensive diagnostic capabilities. Low radiation doses to patients, high reliability, and reasonable cost. Through a digital representation of the image, it becomes possible to achieve quick high-precision diagnostics, reliable archiving, and highly reliable transmission of images over significant distances [3]. However, there exist a number of factors that produce distortions in information in the course of generating information and in its transmission over communication lines.
External disturbances, referred to as interference or noise. Arc the principal reasons why a fluoro turns out to be of poor quality. In terms of manifestation, we distinguish between noise at a frequency up to 40 Hz induced by the movement of the patient in the course of examination (motion artifact); network noise that arises as the powerful medical equipment is connected to the network (pulse noise); and noise due to low-level contacts, communication channels, and external pick-up [4]. The latter types are random in nature, their components arc uncorrected, and are classified as white noise. Effective noise suppression that preserves important de-
tails is a fundamental problem in processing medical images and a basic trend in efforts aimed at improving the systems used to process these images.
While suppressing white and pulse noise, it is also necessary to preserve clarity of the boundaries of objects and fine details which may be comparable in terms of intensity with the noise. This is usually achieved by means of digital linear or wavelet filters. In a number of cases, however, these filters are not able to assure a required level of noise suppression while at the same time preserving the parts of the image.
We will use a method involving decomposition into empirical modes, a method which is an adaptive method of analysis, to achieve effective suppression of noise in medical images. The basis used at decomposition is designed directly from the image (signal), which allows to consider local features of the image, its internal structure and presence of hindrances of a various look [5]. Processing of images by means of such a method produces high-quality noise suppression of noise based on a division of the image into frequency components and the analysts of these components, threshold processing, and inverse recovery without the presence of noise.
List of reference links
1. Catalog of Medical Equipment Manufactured by Nucletron. - URL: http://www.fluro.ukrbiz.net, accessed May 15, 2010.
2. Catalog of Medical Equipment Manufactured by Oldelft. - URL: http://www.medafarm.ru, accessed May 15, 2010.
3. Catalog of Medical Equipment Manufactured by the Institute of Nuclear Physics, Russian Academy of Sciences, Siberian Division. - URL: http://www.equipnet.ru, accessed May 15, 2010.
4. Baru, S. E. Scanners - Their Properties and Status in Modern Digital Fluorography / S. E. Baru. - URL: http://zhuravlev.info/index.php, accessed May 12, 2010.
5. Kharkevich, A. A. The Struggle Against Noise [in Russian] / A. A. Kharkevich. - Moscow : Nauka, 1965.
Тычков Александр Юрьевич кандидат технических наук, директор студенческого научно-производственного бизнес-инкубатора, Пензенский государственный университет
E-mail: tychkov-a@mail.ru
Tychkov Alexsandr Yurievich Candidate of engineering sciences, director of research and production of student business incubator, Penza State University
Чураков Петр Павлович
доктор технических наук, профессор, декан факультета приборостроения, информационных технологий и систем, Пензенский государственный университет E-mail: fait@pnzgu.ru
Churakov Peter Pavlovich
Doctor of engineering sciences, professor,
dean of instrumentation,
information technology and systems,
Penza State University
Кузьмин Андрей Викторович
кандидат технических наук, доцент, кафедра теоретической и прикладной механики,
Пензенский государственный университет
E-mail: flickerlight@inbox.ru
Kuzmin Andrey Viktorovich Candidate of engineering sciences, associate professor, sub-department of theoretical and applied mechanics, Penza State University
УДК 62-503.55 Тычков, А. Ю.
Обработка медицинских изображений / А. Ю. Тычков, П. П. Чураков, А. В. Кузьмин // Модели, системы, сети в экономике, технике, природе и обществе. -2013. - № 1(5). - С. 117-119.
