Noise removal from medical image using fusion technique based on DWT coefficient
A H Russell · Journal of Physics Conference Series · 2019
Abstract This paper introduces new manner for removing noise from medical images using wavelet transform based on fusion technique. The main goal of this research is restaurant the medical image based on peak signal to noise ratio value. This paper contains of three key stages are discrete wavelet transform, denoising wavelet and fusion technique. The new idea is compared each sub band of wavelet with other sub band depend on PSNR value and is selected the higher value of measure. First, perform 2level of discreet wavelet transform on the noisy medical images. Then, perform denoising wavelet techniques (soft and hard thresholding) that achieved by threshold value for detail coefficient and compared it with wavelet coefficients for detail sub band. Second, select sub band that contains the higher peak signal to noise ratio value that has less amount of noise from each image and it is merged to form the fused image. Finally, perform inverse discrete wavelet transform to convert the output image (fused image) from frequency domain to spatial domain. The outcomes of the work showed that soft thresholding is better than hard thresholding based on higher PSNR value. The fusion process gives better results than wavelet denoising technique.