Identification of static objects in the respiratory epithelium

• Published in: 2018 ELEKTRO pp. 1 - 5
• Main Authors: Hargas, Libor, Jabloncik, Frantisek, Koniar, Dusan, Volak, Jozef, Loncova, Zuzana
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01-05-2018
• Subjects: cilia; Classification algorithms; Feature extraction; image segemntation; Image segmentation; local binnary pattern; Mechatronics; Object recognition; Support vector machines; Task analysis
• DOI: 10.1109/ELEKTRO.2018.8398319

The article deals with methods of detection of microscopic objects in video sequences while using the techniques of image processing and segmentation. It is focused on searching for static objects of interest - cilia of respiratory epithelium, with emphasis on elimination of the false-positive detections, i.e. detections of image artefacts. The article describes currently used segmentation techniques in the field of digital image processing, while besides the description of basic principles it also illustrates the application of particular methods into investigated images. The proposed algorithm for identifying objects of interest is described too. Finally, obtained results are analysed mainly from the accuracy point of view, summed up and compared to other so far used techniques.