Naslov
An efficient SA algorithm for brain hemorrhage image analysis

Autori
Lončarić, Sven ; Majcenić, Zoran

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni

Izvornik
Book of abstracts from 1st Croatian international symposium on computer assisted surgery and telesurgery / Klapan, Ivica - Zagreb, 1999

Skup
1st Croatian international symposium on computer assisted surgery and telesurgery

Mjesto i datum
Zagreb, Hrvatska, 30.03.1999. - 02.04.1999

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Domaća recenzija

Ključne riječi
simulated annealing; image segmentation; computed tomography

Sažetak
A high-level analysis of medical images is required for quantitative measurements of various image regions or organs of interest. For example, the study of human spontaneous intracerebral brain hemorrhage (ICH) requires various ICH measurements such as the hemorrhage volume and its change over time. In this paper, a computationally efficient simulated annealing (SA) algorithm is presented that has been tested on the problem of segmentation of CT images containing ICH. The speed-up of the proposed method is based on the use of multiresolution image pyramid representation. The proposed multiresolution simulated annealing (MSA) algorithm starts at a low image resolution therefore having lower computational complexity. The segmentation results achieved at lower resolutions are used as initial approximation for the next step of segmentation which is performed on the next higher resolution. In such a manner it has been demonstrated that it is possible to achieve a five-fold to six-fold reduction in the computation time required for image segmentation of a typical CT image. The method has been tested on CT brain images of patients having ICH. The method is robust and has performed correct segmentation of CT images.

Izvorni jezik
Engleski

Znanstvena područja
Elektrotehnika

POVEZANOST RADA