Avtor/Urednik		Tomaževič, Dejan; Cigale, Boris; Zazula, Damjan; Pernuš, Franjo
Naslov		Primerjava petih filtrov pri razgradnji ultrazvočnih posnetkov jajčnikov z aktivnimi konturami
Tip		članek
Vir		In: Zajc B, editor. Zbornik 8. elektrotehniške in računalniške konference ERK '99. Zvezek B. Računalništvo in informatika, umetna inteligenca, razpoznavanje vzorcev, biomedicinska tehnika, močnostna elektrotehnika, didaktika, študentski članki; 1999 sep 23-25; Portorož. Ljubljana: Slovenska sekcija IEEE, Region 8,
Leto izdaje		1999
Obseg		str. 265-8
Jezik		slo
Abstrakt		Knowledge about the status of the female reproductive system is important for fertility problems and age-related family planning. The volume of these fertility reguests in our emancipated society is steadily increasing. Important information about the ovarian aging, i.e. number of follicles, size, position and response to hormonal stimulation, can be obtained by intravaginal 2D ultrasound imaging of the follicles in the ovary. The manual analysis of a number of follicles is tedious, time consuming, and error-prone. In this paper we used GVF active contours to automatically segment the follicles. Because the ultrasound images are corrupted with speckle noise, which makes segmentation difficult and unrelia6le, prior to the segmentation the images are filtered. Five different filtering algorithms, Koo's filter, Zong's filter, linear diffusion, non-linear diffusion of Perona and Malik, and the morphological closing operator were implemented. We compare the five filters and discuss the segmentation results.
Deskriptorji		OVARY PREGNANCY FOLLICULAR PHASE IMAGE PROCESSING, COMPUTER-ASSISTED