Title: Edge finding in magnetic resonance imaging applications: the calculation of the first order derivative of two dimensional images

Authors: Farouk Yahaya

Addresses: University of Information Science and Technology 'St. Paul the Apostle', Partizanska B.B., Ohrid, 6000 (FYROM), Macedonia

Abstract: This paper proposes a novel edge detection algorithm based on model fitting and the calculation of the first order derivative of two-dimensional images. It is central to mention that the aforementioned calculation can only be done through approximations since the image being used is made of a sequel of discrete samples, and the continuity is brought in through the model function which needs to offer the property of first order differentiability. The cubic bivariate polynomial was used in this paper. Indeed, the partial first order derivatives of the images are calculated after fitting the model function to the image data. It is found that the cubic bivariate polynomial shows visible edges of the images as seen in the results section. The usefulness of the methodology is also tested on a range of images, practical applications, and performance analysis under realistic conditions.

Keywords: cubic bivariate polynomial; edge finding; first-order derivative; motion artefacts; non-invasive; model-fitting; biomedical imaging; magnetic resonance imaging; edge-contour; two-dimensional images; edge detection.

DOI: 10.1504/IJAPR.2017.086595

International Journal of Applied Pattern Recognition, 2017 Vol.4 No.3, pp.226 - 245

Received: 11 Nov 2016
Accepted: 26 Apr 2017
Published online: 12 Sep 2017 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article