An approach for segmentation of X-ray angiographic image based on region-growing and structure inferring.
- Author:
Chuan MEI
;
Guiliang WU
;
Yuan YANG
;
Lan XIE
;
Jiaju HE
;
Shaolin LI
;
Shoujun ZHOU
- Publication Type:Journal Article
- MeSH:
Algorithms;
Coronary Angiography;
Humans;
Radiographic Image Enhancement;
methods;
X-Rays
- From:
Journal of Biomedical Engineering
2014;31(2):413-420
- CountryChina
- Language:Chinese
-
Abstract:
We presented a new method for vessel segmentation and vascular structure recognition for coronary angiographic images. During vessel segmentation, a new vessel function was proposed to attain vessel feature map. Then the region growing algorithm was implemented with an automatic selection of seed point, extraction of main vessel branch, and vessel detail repairing. In the algorithm of vascular structure recognition, a fuzzy operator was used, which can detect the structures of vascular segments, bifurcations, crosses, and tips. The experimental results indicated that there was about 5 percent larger vessel region which was extracted by the proposed segmentation method than that by the simple region growing algorithm, and several thinner vessels were resumed from the lower gray region. The results also indicated that the fuzzy operator could correctly infer the simulative and real vessel structure with 100% and 90.59% correctness rate on the average, respectively.
