Fringe Noise Removal Using Sigmoid Frequency Filtering

The patent changes the filtering approach from stepped band-stop filter to Gaussian filter, modifying the frequency domain parameters to create a smooth transition instead of abrupt cutoff. This parameter change eliminates the ringing artifacts while maintaining noise removal effectiveness

Instead of using a stepped filter that completely removes frequencies in the stop band, the patent inverts the approach by using a Gaussian filter that gradually attenuates frequencies. This inverted filtering strategy prevents the formation of ringing artifacts at image edges

The patent applies local quality by using Gaussian filtering that provides different attenuation levels at different frequencies. The filter preserves useful information near noise frequency points while still suppressing fringe noise, creating a non-uniform but optimized filtering effect across the frequency spectrum

The patent uses partial action by applying Gaussian filtering that doesn't completely eliminate frequencies in the noise band but rather attenuates them partially. This partial denoising approach maintains useful image information while still achieving effective noise suppression