Рақамли технологияларнинг назарий ва амалий масалалари 8-jild 1-son (2025) · 105-113-betlar
Optimization of identification of micro-objects based on morphological and brightness characteristics of image points
Жуманов, И.И., Сафаров, Р.А., Джуманов, О.И.
Annotatsiya
A methodology has been developed for optimizing the identification, recognition, and classification of micro-objects based on dynamic models combined with neural networks with tools for filtering impulse interference and noise, foreign particles, and other defects of image points; the methods and models of which are used in the systems of palynology, ecology, environmental protection, medicine, and other fields of knowledge. Identification mechanisms have been developed, including tools for contour extraction, segmentation, obtaining segment boundaries with hard and soft thresholds, and filtering using morphological characteristics of the image. Estimates of identification errors due to inadequacy of approximation, interpolation, and extrapolation of the image contour have been obtained. A software package for recognizing and classifying images of micro-objects has been developed, in which cubic, biquadratic, and interpolation spline functions and wavelet transformation have been synthesized.
изображениемикрообъектидентификацияраспознаваниеклассификацияфильтрацияэффективностьпрограммный комплексimagemicro-object
Metadata manbasi: jurnal OAI-PMH arxivi · Sindex toʻliq matnni saqlamaydi, manbaga havola beradi.