Image enhancement by using Hybrid methods based on Artificial Neural Networks

In this study, the elimination of zero diffraction order and twin image from three dimensional reconstructed images in digital holography and the image enhancement are desired to achieve in a short time by using hybrid methods. In addition, the brightness of the reconstructed three dimensional image is aimed to increase by using hybrid methods based on artificial neural network for the first time. To calculate the accurate phase of the hologram using in Gerchberg-Saxton algorithm (GSA), the Fourier transform algorithm and 1- dimensional continuous wavelet transform (1D-CWT) are used. These methods are expressed as hybrid methods. The usage of 1D-CWT for using in GSA is a first attempt. The intended purpose of using 1D-CWT is to reduce the twin image and zero diffraction order with performing minimum iteration numbers. The ratio of image enhancement is given by calculating the normalized root mean square values.