Line feature enhancement using a new shift-invariant non-aliasing ridgelet transform

  • Authors:

  • He Yan;Youjia Fu;Guangzhi Yin

  • Affiliations:

  • Key Lab. for the Exploitation of Southwestern Resources & the Environmental Disaster Control Engineering, Ministry of Education, Chongqing, China and College of Computer Science, Chongqing Uni ...;College of Computer Science, Chongqing University of Technology, Chongqing, China;Key Lab. for the Exploitation of Southwestern Resources & the Environmental Disaster Control Engineering, Ministry of Education, Chongqing, China

  • Venue:
  • Transactions on Edutainment VII
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

A new shift-invariant non-aliasing Ridgelet transform was presented to avoid aliasing and shift-variant in old Ridgelet transform. Firstly, image size was expanded twice via interpolating zero values,so that the sampling rate in Radon transform domain was improved and the aliasing was reduced; Secondly, the one-dimensional discrete wavelet transform in the old Ridgelet transform was replaced by one-dimensional dual-tree Complex wavelet transform(DCWT) which had both shift-invariant and double directional analysis ability. On this basis, a new image enhancement method for linear features was proposed by using zero-mean Gaussian mode in the shift-invariant non-aliasing Ridgelet transform domain. Experimental result shows that the presented Ridgelet transform can avoid effectively the ‘scratch around'phenomenon in reconstructed image and the visual effects of the image enhancement is improved obviously.