Image fusion of multidirectional wavelet transforms for image denoising

• Main Authors: Ochotorena, Carlo Noel E., Ochoterena, Cecille Adrianne, Sybingco, Edwin
• Format: text
• Published: Animo Repository 2014
• Subjects: Wavelets; Image systems; Engineering
• Online Access: https://animorepository.dlsu.edu.ph/faculty_research/10822
• Institution: De La Salle University

Image denoising in the wavelet domain has been attractive to researchers in the past decade due to its suitable properties which lead to a smooth denoised image. However, limitations in edge representation have been found particularly with diagonal edges which may be destroyed in the denoising process. Direction-sensitive variants of the transform have been proposed by various researchers but they are often computationally expensive and difficult to implement in an actual system. This work presents a direction-sensitive technique of denoising based on a reversible rotation process combined with a simple discrete wavelet transform and image fusion methods. Test results on the said methods show significant gains over the traditional wavelet denoising process. Additionally, the simple and non-adaptive nature of the process makes it attractive for software and hardware implementation.