A Novel Algorithm for Denoising using Adaptive Thresholding Based Dual Tree Complex Wavelet Transform (DTCWT) on Ultrasound Medical Image

Abstract

Noise removal, a crucial step is in systems of digital image processing. Though many algorithms are derived for noise removal there is no single algorithm for all types of noise removal. Discrete wavelet transformation has some disadvantages such as it is computationally intensive, shift variance etc. to resolve this challenge, a research proposal  method is designed using dual complex wavelet transform algorithm which is far better than the old methods and produces good results. DWT gives extend to Dual-tree complex wavelet transform (DTCWT). Computation of every signal is done along DTCWT complex transform. Each wavelet implements a threshold value, the coefficient value that are greater than threshold value of the coefficients are stored (Lesser are ignored). Then it is passed to many filters to remove different types of noises. Finally an enhanced image with reduced noise will be obtained.  Image with noise could be improvised in quality of visual; it simply changes the coefficients with the help of soft-thresholding method. Noises of additive , Speckle, multiplicative and Gaussian and its factors affect images in ultrasound, which minimize the image quality and effects the human interpretation. So, DTCWT based thresholding approach helps to minimize the noise rate considerably for the provided US image. The experimental result confirms that the proposed approach provides better performance with respect to higher PSNR, SSIM and lower MSE, execution time rather than the previous Fisz transformation and DWT methods.