Wyniki wyszukiwania - Biblioteka Nauki

Nowa wersja platformy, zawierająca wyłącznie zasoby pełnotekstowe, jest już dostępna.
Przejdź na https://bibliotekanauki.pl

Ograniczanie wyników

Znaleziono wyników: 2

Liczba wyników na stronie

Wyniki wyszukiwania

Wyszukiwano:
w słowach kluczowych: thresholding methods

Sortuj według:

Ogranicz wyniki do:

Detection of ceramic insulators using color thresholding

100%

Osuchowski J.

2017

tom z. 75

59--60

The purpose of this work is to outline the problems with insulators detection with computer vision and image processing methods. A basic power line insulator detection algorithm is presented. The algorithm used a HSV (Hue, Saturation, Intensity) color model, color thresholding and morphological erosion for image filtering. An important element of the work is the analysis of the reliability of the overhead line insulators and the impact of their failure on the line maintenance.

Thresholding techniques for segmentation of atherosclerotic plaque and lumen areas in vascular arteries

75%

Markiewicz T. , Dziekiewicz M. , Osowski S. , Maruszyński M. , Kozłowski W. , Bogusławska-Walecka R.

tom Vol. 63, nr 1

269--280

The paper develops the automatic methods of segmentation of the blood vessel area in the images of the multi-slice computed tomography, allowing to separate the lumen from the atherosclerotic plaque areas. The solution is based on the application of different implementations of thresholding, including between class variance in a bimodal mode, Gaussian mixture modeling, clustering technique, polynomial and multilayer perceptron approximations. These methods are compared with many examples of arteries of different percentage of the plaque occupancy in the iliac and femoral arteries. The numerical results of segmentation have been verified by the medical experts and prove its usefulness in medical practice. The presented system can find application in an automatic evaluation of the atherosclerosis progression/regression of patients on the basis of sequence of Computed Tomography slice images.