Filters

Toolbar:

Menu:    Processing > Filters…

Use this command to select one of ImageWarp's or user-defined spatial filters and apply it to an image. Most of the filters are mathematical functions that assign values to pixels based on the values of neighboring pixels. This is often described as kernel operations or convolution. The kernel is a matrix of coefficients, which are multiplied by each pixel and its neighbors and then summed to produce a resultant value. The result is placed in the output image at the location of the original pixel. This operation is applied to all pixels in the input image by moving the kernel across the image pixel by pixel. Some filters are not based on convolution, but they still generate a value for the target pixel by sampling the neighbors and applying some other mathematical function to them.

A special condition may arise for boundary pixels that have no neighbors on one side. ImageWarp resolves this situation by utilizing the hidden borders (overscan) that surround each image and allow the same kernel operators to be applied to all pixels in the image without checking the boundary conditions. That significantly increases the speed of processing, but limits the maximum size of the kernel matrix to the overscan size multiplied by two. Use the Overscan option in the Preferences to adjust the border size for all images.

If the input image is a color one, ImageWarp can perform filtering in certain color space, defined by the Processing space option in the Preferences If the color space for processing differs from the input image type, ImageWarp converts the image into the corresponding internal type, processes it and then converts it back into the original type.

Spatial filters fall into one of four general categories: image enhancement, noise reduction, edge and texture extraction and advanced operations. Each category is presented in a separate tab of the Filters functional dialog box. Select one of the following functions:

Enhance

Click… To… Emphasize emphasize fine detail in an image by applying the unsharp mask. Sharpen sharpen an image by applying the sharpening filter. Clarify enhance an image by applying the nonlinear lowpass filter. MaxMin emphasize edges of objects while suppressing the noise. Normalize enhance the contrast of an image by normalizing its histogram. Equalize enhance the contrast of an image by equalizing its histogram.

Smooth

Click… To… Lowpass blur noise in an image by applying the lowpass filter. Median remove impulse noise from an image by applying the median filter. Sigma enhance an image by applying a non-linear smoothing. Gauss suppress noise in an image by applying the Gauss filter. Salt and Pepper remove salt & pepper type of noise from an image Deinterlace remove motion artifacts from an interlaced image.

Edges

Click… To… Sobel apply the Sobel operator to an image. Prewitt apply the Prewitt operator to an image. Laplace apply the Laplacian operator to an image . Roberts apply the Roberts operator to an image. HighPass extract high frequency information from an image. Gradient extract gradient information from an image. Range apply the range operator to an image. Variance apply the variance operator to an image.

Advance

Click… To… Emboss create a 3D appearance of an image. Floyd apply the Floyd-Steinberg dithering to an image. Pixelate create a mosaic effect on an image. Fourier Applies the Fourier transform to an image Convolve perform a linear convolution on an image.