Reducción de Ruido en Señales de Espectroscopia en Resonancia Magnética con Transformada Wavelets COINFLET Y BIORTHOGONAL

Abstract

This document presents the application of the Coinflet Wavelet Transform and the Biorthogonal wavelet family to reduce noise in magnetic resonance images, used in non-invasive medical processes. Noisy measured signals are processed by a noise cleaning algorithm in Matlab; then the signals are converted to the frequency domain where they are processed Compressed and cleaned from Noisy source. The wavelet transform is a technique used to manipulate, analyze, and compress signals more efficiently in magnetic resonance imaging (MRI) noise reduction applications, which can often be affected by Gaussian noise from from different sources, said Gaussian noise is difficult to clean due to the amount of information that must be processed in diagnoses of ligament ruptures to tumors. The Coiflet wavelet uses compact support techniques and improved regularity properties to capture fine details in the image and preserve structural features. relevant, while reducing unwanted noise. In contrast, the Biorthogonales family of wavelets have the advantage of having separable analysis and synthesis filters, which simplifies the filtering process and improves efficiency. The noise reduction with wavelet presented satisfactory results at high frequencies where the Coinflet wavelet stood out in comparison to the Daubechies and the Biorthogonal using the Penallo Threshold technique.