Search academic texts

Information × Registration Number 0223U001239, 0121U112176 , R & D reports Title Methods of multichannel image processing in mobile systems based on prediction and machine learning popup.stage_title Head Lukin Volodymyr V., Доктор технічних наук Registration Date 26-01-2023 Organization National Aerospace University "Kharkiv Aviation Institute" popup.description2 The report presents the main results of the development of methods and fast algorithms, as well as state-of-the-art multichannel image processing tools based on the assessment of complexity and data quality for processing steps such as lossy compression and classification. It is shown that atomic transformations can be an effective means of compression, for which the probabilities of correct classification at different degrees of compression are estimated. It is also proven that the complexity of the image significantly affects the efficiency of its processing at almost all stages, the indicators of complexity are discussed. Various classifiers have been analyzed, the effectiveness of classifiers based on trained neural networks has been proven, and various learning strategies have been considered. The features of lossy compression of images distorted by interference were studied. Adaptive approaches based on simultaneous evaluation of image characteristics and noise are proposed. Part of the developed methods and algorithms is implemented in the form of software-algorithmic support for processing color images, multi-channel remote sensing data, medical images, etc. Product Description popup.authors Abramov Sergіy K Abramova Victoria V. Vasilyeva Irina K. Kovalenko Bogdan V Krivenko Sergey S. Lі Fanfan --- Makarichev Victor O. Naumenko Victoriia V. Proskura Galina A. Rebrov Volodymyr Rubel Andriy S. Rubel Oleksiy S Uss Mykhailo L. Chernova Galyna A. popup.nrat_date 2023-01-26 Close