Fast adaptive learning algorithm for sub-band adaptive thresholding function in image denoising
by G.G. Bhutada, R.S. Anand, S.C. Saxena
International Journal of Computational Intelligence Studies (IJCISTUDIES), Vol. 1, No. 3, 2010

Abstract: The speed of image denoising by adaptive thresholding approach in Wavelet Transform (WT) domain depends mainly upon the learning algorithm used for optimising the performance of adaptive thresholding function. In this context, in the literature, steepest gradient-based optimisation technique has been used in WT-based thresholding neural network (WT-TNN) approach, which has low learning speed. In this paper, a new computationally efficient approach, that is, Particle Swarm Optimisation (PSO)-based approach has been proposed in place of steepest gradient-based approach. The proposed hybrid computing approach utilises the features of WT-TNN approach and enhances the speed of optimisation by PSO technique. It also yields better performance of denoising as compared to WT-TNN approach. In the proposed approach, crucial problem of initialisation of thresholding parameters gets automatically sorted out besides learning time becoming independent of noise level of the image. The proposed approach also enhances edge preservation, when implemented with bior6.8 wavelet filters.

Online publication date: Thu, 26-Aug-2010

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Computational Intelligence Studies (IJCISTUDIES):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com