Image Denoising Based Improved Curvelet Threshold

  • Ahmed A. Hashimϯ †university of information technology and communications, department of media technology engineering, , Iraq-baghdad
Keywords: Peak signal to noise ration PSNR, Mean square error MSE., wavelet transform, ridgelet transform, Image denoising, curvelet transform



Although wavelets are powerful as a tool in image processing it has three serious drawbacks: shift sensitivity, poor directionality and lack of phase information. Wavelets suitable for dealing with objects with point singularities; it can only capture limited directional information due to its poor orientation selectivity. Through decomposing the image into a series of high-pass and low-pass filter bands, the wavelet transform extracts directional information that capture vertical, horizontal, and diagonal activity. However, in noisy images, these three linear directions are limiting and might not capture enough directional information, like medical CT scans, which do not have strong vertical, horizontal, or diagonal directional elements.

Ridgelet transform improves Multiresolution Analysis MRA segmentation; however, they capture structural information of an image based on multiple radial directions in the frequency domain. Line singularities in ridgelet transform provides better edge detection than its wavelet counterpart. One limitation to use ridgelet in image segmentation is that ridgelet is most effective in detecting linear radial structures, which are not dominant in images. The curvelet transform is a recent extension of ridgelet transform that overcome ridgelet weaknesses. Curvelet is proven to be particularly effective at detecting image activity along curves instead of radial directions which are the most comprising objects of images. However, the fact that at sufficiently fine scales, a curved edge is almost straight, and so to capture curved edges, one ought to be able to deploy ridgelets in a localized manner, at sufficiently fine scales.

In this paper, a new method is used combining the Window Shrink threshold and Bayes Shrink threshold based on Curvelet transform to enhance removing Additive Weight Gaussian Noise AWGN noise from image. It has better PSNR than the traditional Curvelet that uses each threshold alone, and the images gotten by this method is better and outperform that of the traditional curvelet and wavelet methods and the results reported here are promising