Smoothing And Denoising
This group contains algorithms for removing noise from images.
- BoxFilter2d: Smooths an image using a box kernel.
- BoxFilter3d: Smooths an image using a box kernel.
- MedianFilter2d: Applies a median operator, which is a non-linear smoothing filter, on a two-dimensional image.
- MedianFilter3d: Applies a median operator, which is a non-linear smoothing filter, on a three-dimensional image.
- GaussianFilter2d: Applies a two-dimensional Gaussian filter using either a separable finite kernel, or a recursive algorithm.
- GaussianFilter3d: Applies a three-dimensional Gaussian filter using either a separable finite kernel, or a recursive algorithm.
- ExponentialFilter2d: Smooths an image using a kernel based on an exponential distribution.
- RecursiveExponentialFilter2d: Smooths a two-dimensional image with a recursive algorithm implementing an exponential filter.
- RecursiveExponentialFilter3d: Smooths a three-dimensional image with a recursive algorithm implementing an exponential filter.
- NonLocalMeansFilter2d: Adaptive filter denoising a two-dimensional image.
- NonLocalMeansFilter3d: Adaptive filter denoising a three-dimensional image.
- CurvatureDrivenDiffusion: Smooths a two-dimensional image using an advanced local edge analysis technique.
- BilateralFilter2d: Applies a two-dimensional edge-preserving smoothing filter.
- BilateralFilter3d: Applies a three-dimensional edge-preserving smoothing filter.
- MajorityFilter2d: Replaces pixels of a two-dimensional image by the most represented value inside their neighborhood.
- MajorityFilter3d: Replaces voxels of a three-dimensional image by the most represented value inside their neighborhood.
- SigmaFilter2d: Performs an adaptive smoothing of a two-dimensional image by excluding any aberrant pixels of a local averaging.
- SigmaFilter3d: Performs an adaptive smoothing of a three-dimensional image by excluding any aberrant voxels of a local averaging.
- Despeckle2d: Smooths a two-dimensional image by replacing any aberrant pixel values by their neighbors mean value.
- Despeckle3d: Smooths a three-dimensional image by replacing any aberrant voxel values by their neighbors mean value.
- NagaoFilter2d: Performs an edge-preserving smoothing of a two-dimensional image by selecting a mean value from different neighborhood configurations.
- NagaoFilter3d: Performs an edge-preserving smoothing of a three-dimensional image by selecting a mean value from different neighborhood configurations.
- SnnFilter2d: Performs an edge-preserving smoothing of a two-dimensional image with the Symmetric Nearest Neighbor filter.
- SnnFilter3d: Performs an edge-preserving smoothing of a three-dimensional image with the Symmetric Nearest Neighbor filter.
- FlowInpainting: Reconstructs missing areas of an image with an inpainting algorithm.
The Smoothing And Denoising group contains:
- Lowpass filters that reduce the contrast and soften the edges of objects in an image. A lowpass filter lets low frequencies go through but attenuates high frequencies and noise. It reduces contrast, but also tends to defocus the image.
- Adaptive filters that self-adjust their transfer function according to an optimizing algorithm, contrary to non-adaptive filters which have static filter coefficients applied on a static filter window. Adaptive filters reduce noise while preserving object edges.