Abstract
Wavelet transformation is one of the most frequent procedures for data denoising, smoothing, decomposition, features extraction, and further related tasks. In order to perform such tasks, we need to select appropriate wavelet settings, including particular wavelet, decomposition level and other parameters, which form the wavelet transformation outputs. Selection of such parameters is a challenging area due to absence of versatile recommendation tools for suitable wavelet settings. In this paper, we propose a versatile recommendation system for prediction of suitable wavelet selection for data smoothing. The proposed system is aimed to generate spatial response matrix for selected wavelets and the decomposition levels. Such response enables the mapping of selected evaluation parameters, determining the efficacy of wavelet settings. The proposed system also enables tracking the dynamical noise influence in the context of Wavelet efficacy by using volumetric response. We provide testing on computed tomography (CT) and magnetic resonance (MR) image data and EMG signals mostly of musculoskeletal system to objectivise system usability for clinical data processing. The experimental testing is done by using evaluation parameters such is MSE (Mean Squared Error), ED (Euclidean distance) and Corr (Correlation index). We also provide the statistical analysis of the results based on Mann-Whitney test, which points out on statistically significant differences for individual Wavelets for the data corrupted with Salt and Pepper and Gaussian noise.
Highlights
As far as it is commonly known, as the result of the physical limitations of most medical systems, signals and images have the tendency to manifest some random noise within signal and image acquisition [1]
The construction of the results based on the spatial modelling is shown in Figure 8 is illustrating the result of achievable correlation level for two specific amounts of noise
From the results it is obvious that for the chosen input images the influence of the used wavelet family is rather low. This correspond to the results shown in Reference [58], where it is mentioned that the Symlet wavelet family outperforms the other families, even though the plotted results do not show a significant difference between the Daubechies and Symlet wavelet families shown in this paper
Summary
As far as it is commonly known, as the result of the physical limitations of most medical systems, signals and images have the tendency to manifest some random noise within signal and image acquisition [1]. We are overcrowded by plenty acquiring medical devices, intended for data acquisition, where the resulting signals and images may be affected by different types of the noise. Considering different nature of individual types of the noise, we can deduce that individual smoothing methods may be differently sensitive and robust to each of the noise type. This hypothesis deals with the fact that there is not unified method, with unified settings which would perform the data smoothing, corrupted by different noise manifestation with the same effectivity [10,11]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
