boundary complexity
Recently Published Documents


TOTAL DOCUMENTS

8
(FIVE YEARS 0)

H-INDEX

2
(FIVE YEARS 0)

2020 ◽  
Vol 17 (5) ◽  
pp. 2007-2013
Author(s):  
Kandavalli Michael Angelo ◽  
S. Abraham Lincon

Identifying and separating objects within an image is a significant challenge due to high object and background variability. This can be obtained through feature extraction approach. There are different ways of extracting the image features. It is based on texture, shape and colour. The present paper aims to study and analyse the various approaches for feature extraction and object recognition. This study proposed a hybrid approach, which is a combination of enhanced Fractal Texture Analysis with Layout Descriptor to overcome the obstacles in image segmentation. It is used to lessen the boundary complexity of the segmented image. First, the image is preprocessed to discard the noise and to retain the adequate details of the image in a perfect way through Adaptive Switching Median Filter. Secondly, it improves the power of the edges detected through a noise-protected edge detector. Finally, it is applied with morphological gradient technique that is a twin function of both shape and texture gradient removal for extorting the qualities of the image. In this way, the proposed methodology directly performs on the colour image which supports to enhance prediction accuracy of the object in terms of colour characteristics that offers better results than the grayscale conversion approach. Moreover, the shape feature is extracted from the preprocessed image depending on the details like compactness, rectangularity, eccentricity and moment invariants.


2013 ◽  
Vol 24 (5) ◽  
pp. 985-993
Author(s):  
Hamid Reza Ghaffari ◽  
Hadi Sadoghi Yazdi

2002 ◽  
Vol 31 (6) ◽  
pp. 1745-1760 ◽  
Author(s):  
János Pach ◽  
Gábor Tardos
Keyword(s):  

Sign in / Sign up

Export Citation Format

Share Document