VSE knjižnice (vzajemna bibliografsko-kataložna baza podatkov COBIB.SI)
  • Computer-assisted detection of ovarian follicles based on ultrasound images
    Potočnik, Božidar ; Viher, Bogdan ; Zazula, Damjan
    The paper deals with the problem of processing and interpretation of clinically recorded ultrasonic images for the reason of following the growth of dominant ovarian folicles in a day-to-day manner. ... The procedure has been conducted by the physicians involved in human reproduction and fertilisation "in vitro" manually, so far with a rather stressing burden of a huge number of readings to be done regularly. We propose three different automatic computer-based identification algorithms. All of them begin with a despecle-filtering phase. Then, the first one is composed of conventional image-processing approaches, with edge processing and region segmentation, in order to estimate the ovary location and afterwards, recognised the dark regions as possible follicles. Two additional approaches are based on cellular automata theory and on cellular neural networks. They represent quite novel approaches in the field of image recognition and the results obtained outperform the conventional algorithms in the sense of sensitivity and specificity. With the former ones go up to 87 % of recognition rate, the latter one increase it to over 90 % and decrease false rate below 12 % at the same time.
    Vir: A számítástechnika orvosi és biológiai alkalmazásai (Str. 24-34)
    Vrsta gradiva - prispevek na konferenci
    Leto - 1998
    Jezik - angleški
    COBISS.SI-ID - 4064278

Vnos na polico