Construction of a novel convolution based fractional derivative mask for image edge analysis
No Thumbnail Available
Date
2016-08
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
KNUST
Abstract
This thesis presents a new approach in constructing a more efficient fractional
derivative mask for image edge analysis based on the definition and properties of
convolution. By the definition of convolution, the generalised Strivastiva-Owa’s
operator was rewritten with its order restricted to the Riemann-Liouville fractional
derivative. Applying linearity, commutative and derivative properties of convolution
to the resultant expression, a new mask with higher efficiency, memory effect and
computational equivalence to the classical edge detector is developed as per the
experimental results obtained. From the experimental results, it is observed that, the
new mask has the potency to find edges in details quite significantly as well as hidden
edges which is a deficiency of the classical edge detectors. It can also be used on a
region growing algorithm during region segmentation acting as an edge function in
its termination process. The experiments conducted on the mask were done using
some selected well known synthetic and medical images with realistic geometry.
Using visual perception and performing both mean square error and peak signal-to noise ratios analysis, the method demonstrated significant advantages over other
known methods