Due to various causes noise is most often present on raw images. Generating masks or detecting edges using the noisy raw image is not satisfactory. The image is to be smoothed by a preprocess.
Gaussian blur is an image smoothing operation done by use of seperable symmetric matrix of convolution coefficients. The size can be any odd nimber between 3 (for a 3x3 grid size) to 11 (for a 11x11 grid). Given the size and standard deviation, the convolution weights are computed and applied.. In YUY2 and YV12 formats only the Y values are smoothed. However if u and or v are made true, those corresponding planes are also smoothed. In RGB formats all 3 color planes are smoothed.
RGB32, RGB24, YUY2 and Planar formats are supported.
|Input clip to be smoothed||clip||none|
|size of grid to use||size||integer||odd number between 3 to 11||3|
|standard deviation||sd||float||more than 0||1.5|
|is u plane smoothing required?||u||boolean||true for yes, false for no||false|
|is v plane smoothing required?||v||boolean||true for yes, false for no||false|
|To Colorit Synopsis|
|To my index page|
|down load plugin|