Feb 5, 2001  Noise  Common types of noise found in images:  1. Uniform (white) noise: 2. Gaussian noise: where m is the mean and σ is the standard deviation.   3. Negative exponential noise: 4. Salt-and-Pepper noise:   Spatial Filters   Convolution of an NxM f(x,y) image with a filter impulse response h(x,y) defined over the set of elements given by H is given by: where 0 ≤ x-i < N    and  0 ≤ y-j < M   Low Pass Filters   1.  Mean (or average) filter (3x3):   This filter works fairly well for Gaussian noise, is poor for salt-and-pepper noise, and has the side-effect of reducing high frequencies.  It is basically a low-pass filter.  The larger the size of the filter, the more pronounced its effects.   2.  Gaussian filter with a = 3.32:  The values of the above filter are calculated from The constants in front are the normalizing constants to make the sum of all the coefficients equal to unity.   High-Pass Filters   It is more difficult to get a high-pass filter than a low-pass filter.  Indeed, to get a high-pass filter, the general procedure is to apply a low-pass filter to the original image and then subtract this low-frequency image from the original image.  The result is then an image containing only high frequencies.  Sometimes it is desired to enhance the high frequencies without removing the low frequencies.  This is called giving the image a high-frequency boost.  This can be done by using a gain factor G (usually between 0 and 2) and following the procedure involved in:  fh(x,y) = f(x,y) + G • [f(x,y) - fmean(x,y)] = (1 + G)•f(x,y) - G•fmean(x,y)  Example of High-Pass filter with boost: start with identity filter:                       after subtracting a mean filter:     add gain G (range 0 to2):           Rank Order Filtering Median filter: This type of filter is much better at removing outlier type noise such as salt-and-pepper and does not reduce high frequencies as much as the mean filter.  It does not use convolution, but instead is one of a class of filters called rank filters, that ranks the values of the neighbors of a specified pixel, and inserts the median value of the neighboring pixels at the location of the specified pixel.    Example: Original image with sharp edge and one outlier:   Image after filtering with a median filter:   Image after filtering with a mean filter:    Derivative Filters Averaging pixels blurs detail in an image, since averaging filters are basically low-pass filters.  This averaging is similar to integration where everything is just summed up.  On the other hand, if it is desired to enhance rapid change (high frequencies) rather than slow change (low frequencies), some sort of differentiation operation must be developed.  The gradient function is used to determine the direction of change in a vector.  For a two-dimensional image, this could be expressed in vector form as:  The magnitude of the gradient is   In terms of images, consider the pixels under a 3x3 mask: To get derivatives in the x and y directions at z5, one simple method is to take differences such as (z6 - z5), to ignore the squaring and root-taking of the definition of the magnitude of the gradient, and simply to approximate However, using masks of even size 2x2 is awkward, so more commonly a 3x3 neighborhood is used, giving This approximation is expressed in the following two masks for the x and y directions, respectively, which are called Prewitt operators:   Another common approximation to the magnitude of the gradient is given by the Sobel operators for x and y directions, respectively:   All of these derivative operators accentuate rapid change (edges) in the image and attenuate large objects.  They are frequently used as edge detectors.   Last modified on February 13, 2001