PERBANDINGAN LEVEL 1 DAN 2 TRANSFORMASI WAVELET DISKRIT UNTUK DENOISING CITRA

Abstract

The wavelet transform is an improvement of the Fourier transform. The Fourier transform changes the domain from the signal domain to the frequency domain, while the wavelet transform changes the signal domain to the scale (dilation) and translation domains. Discrete wavelet transformation (DWT) can be used to reduce noise in images. In this study, level 1 and 2 DWT methods for image denoising are compared. Lena and Mandrill grayscale test images of 512×512 pixels were used. The wavelets used are Haar, symlets, biorthogonal, coiflets, and Daubechies wavelets. We compare the PSNR value and computation time for Lena and Mandrill images using DWT level 1 and 2. From the research results, the highest PSNR values are for DWT level 2. As for the fastest computation times, they are for DWT level 1.