Publication detail

Artery Wall Detection Based on Image Profile Classification

OMRAN, Y. ŘÍHA, K. DUTTA, M.

Original Title

Artery Wall Detection Based on Image Profile Classification

English Title

Artery Wall Detection Based on Image Profile Classification

Type

conference paper

Language

en

Original Abstract

In this article we propose a new method for the accurate locating of points on the arterial wall in ultrasound images. The system takes as an input an ultrasound image that contains the common carotid artery, this image is the outcome of an object detector. The proposed method suggests a way to detect points on the arterial wall in the transverse section of the common carotid artery, because knowing the locus of these points is very essential in constructing the initial contour for a more detailed segmentation using Active contour algorithms. The detection of a circular object may seem a classic task in image processing, but ultrasound images have special characteristics that makes it a harder task which needs a different type of processing. The system uses a Support Vector Machine classifier to classify image profiles in different positions in the image, in order to recognize the points on the arterial wall. This is an initial step towards the analysis of cardiac diseases, or for tracking the movement of CCA.

English abstract

In this article we propose a new method for the accurate locating of points on the arterial wall in ultrasound images. The system takes as an input an ultrasound image that contains the common carotid artery, this image is the outcome of an object detector. The proposed method suggests a way to detect points on the arterial wall in the transverse section of the common carotid artery, because knowing the locus of these points is very essential in constructing the initial contour for a more detailed segmentation using Active contour algorithms. The detection of a circular object may seem a classic task in image processing, but ultrasound images have special characteristics that makes it a harder task which needs a different type of processing. The system uses a Support Vector Machine classifier to classify image profiles in different positions in the image, in order to recognize the points on the arterial wall. This is an initial step towards the analysis of cardiac diseases, or for tracking the movement of CCA.

Keywords

SVM;Ultrasound imaging;Segmentation;Carotid artery

RIV year

2013

Released

11.09.2013

Location

Senec

ISBN

978-80-227-4026-5

Book

Proceedings of 15th International Conference on Research in Telecommunication Technologies

Pages from

64

Pages to

67

Pages count

4

Documents

BibTex


@inproceedings{BUT101534,
  author="Yara {Omran} and Kamil {Říha} and Malay Kishore {Dutta}",
  title="Artery Wall Detection Based on Image Profile Classification",
  annote="In this article we propose a new method for the accurate locating of points on the arterial wall in ultrasound images. The system takes as an input an ultrasound image that contains the common carotid artery, this image is the outcome of an object detector. The proposed method suggests a way to detect points on the arterial wall in the transverse section of the common carotid artery, because knowing the locus of these points is very essential in constructing the initial contour for a more detailed segmentation using Active contour algorithms. The detection of a circular object may seem a classic task in image processing,
but ultrasound images have special characteristics that makes it a harder task which needs a different type of processing. The system uses a Support Vector Machine classifier to classify image profiles in different positions in the image, in order to recognize the points on the arterial wall. This is an initial step towards the analysis of cardiac diseases, or for tracking the movement of CCA.",
  booktitle="Proceedings of 15th International Conference on Research in Telecommunication Technologies",
  chapter="101534",
  howpublished="electronic, physical medium",
  year="2013",
  month="september",
  pages="64--67",
  type="conference paper"
}