{
    "mode": "man",
    "parameter": "pnmnlfilt",
    "section": "1",
    "url": "https://www.chedong.com/phpMan.php/man/pnmnlfilt/1/json",
    "generated": "2026-06-16T06:21:19Z",
    "synopsis": "pnmnlfilt alpha radius [pnmfile]",
    "sections": {
        "NAME": {
            "content": "pnmnlfilt  -  non-linear filters: smooth, alpha trim mean, optimal estimation smoothing, edge\nenhancement.\n",
            "subsections": []
        },
        "SYNOPSIS": {
            "content": "pnmnlfilt alpha radius [pnmfile]\n\n",
            "subsections": []
        },
        "DESCRIPTION": {
            "content": "pnmnlfilt produces an output image where the pixels are a summary of multiple pixels near the\ncorresponding location in an input image.\n\nThis program works on multi-image streams.\n\nThis  is something of a swiss army knife filter. It has 3 distinct operating modes. In all of\nthe modes each pixel in the image is examined and processed according to it and its surround‐\ning  pixels  values.  Rather than using the 9 pixels in a 3x3 block, 7 hexagonal area samples\nare taken, the size of the hexagons being controlled by the radius parameter. A radius  value\nof  0.3333 means that the 7 hexagons exactly fit into the center pixel (ie.  there will be no\nfiltering effect). A radius value of 1.0 means that the 7 hexagons exactly fit  a  3x3  pixel\narray.\n",
            "subsections": [
                {
                    "name": "Alpha trimmed mean filter.    (0.0 <= alpha <= 0.5)",
                    "content": "The value of the center pixel will be replaced by the mean of the 7 hexagon values, but the 7\nvalues are sorted by size and the top and bottom alpha portion of the 7 are excluded from the\nmean.  This implies that an alpha value of 0.0 gives the same sort of output as a normal con‐\nvolution (ie. averaging or smoothing filter), where radius will determine the  \"strength\"  of\nthe filter. A good value to start from for subtle filtering is alpha = 0.0, radius = 0.55 For\na more blatant effect, try alpha 0.0 and radius 1.0\n\nAn alpha value of 0.5 will cause the median value of the 7 hexagons to be used to replace the\ncenter  pixel  value. This sort of filter is good for eliminating \"pop\" or single pixel noise\nfrom an image without spreading the noise out or smudging features on  the  image.  Judicious\nuse  of  the radius parameter will fine tune the filtering. Intermediate values of alpha give\neffects somewhere between smoothing and \"pop\"  noise  reduction.  For  subtle  filtering  try\nstarting with values of alpha = 0.4, radius = 0.6  For a more blatant effect try alpha = 0.5,\nradius = 1.0\n"
                },
                {
                    "name": "Optimal estimation smoothing. (1.0 <= alpha <= 2.0)",
                    "content": "This type of filter applies a smoothing filter adaptively over the image.  For each pixel the\nvariance of the surrounding hexagon values is calculated, and the amount of smoothing is made\ninversely proportional to it. The idea is that if the variance is small then  it  is  due  to\nnoise in the image, while if the variance is large, it is because of \"wanted\" image features.\nAs usual the radius parameter controls the effective radius, but  it  probably  advisable  to\nleave  the radius between 0.8 and 1.0 for the variance calculation to be meaningful.  The al‐\npha parameter sets the noise threshold, over which less smoothing will be done.   This  means\nthat  small  values  of  alpha will give the most subtle filtering effect, while large values\nwill tend to smooth all parts of the image. You could start with values like alpha = 1.2, ra‐\ndius  =  1.0  and try increasing or decreasing the alpha parameter to get the desired effect.\nThis type of filter is best for filtering out dithering noise in both bitmap  and  color  im‐\nages.\n"
                },
                {
                    "name": "Edge enhancement. (-0.1 >= alpha >= -0.9)",
                    "content": "This is the opposite type of filter to the smoothing filter. It enhances edges. The alpha pa‐\nrameter controls the amount of edge enhancement, from subtle (-0.1) to  blatant  (-0.9).  The\nradius  parameter  controls  the effective radius as usual, but useful values are between 0.5\nand 0.9. Try starting with values of alpha = 0.3, radius = 0.8\n"
                },
                {
                    "name": "Combination use.",
                    "content": "The various modes of pnmnlfilt can be used one after the other to get the desired result. For\ninstance  to turn a monochrome dithered image into a grayscale image you could try one or two\npasses of the smoothing filter, followed by a pass of the  optimal  estimation  filter,  then\nsome  subtle  edge  enhancement. Note that using edge enhancement is only likely to be useful\nafter one of the non-linear filters (alpha trimmed mean or  optimal  estimation  filter),  as\nedge enhancement is the direct opposite of smoothing.\n\nFor  reducing  color  quantization  noise  in images (ie. turning .gif files back into 24 bit\nfiles) you could try a pass of the optimal estimation filter (alpha 1.2, radius 1.0), a  pass\nof  the  median  filter (alpha 0.5, radius 0.55), and possibly a pass of the edge enhancement\nfilter.  Several passes of the optimal estimation filter with declining alpha values are more\neffective than a single pass with a large alpha value.  As usual, there is a tradeoff between\nfiltering effectiveness and loosing detail. Experimentation is encouraged.\n\nReferences:\nThe alpha-trimmed mean filter is based on the description in IEEE CG&A May 1990  Page  23  by\nMark  E.  Lee  and Richard A. Redner, and has been enhanced to allow continuous alpha adjust‐\nment.\n\nThe optimal estimation filter is taken from an article \"Converting Dithered  Images  Back  to\nGray  Scale\"  by Allen Stenger, Dr Dobb's Journal, November 1992, and this article references\n\"Digital Image Enhancement and Noise Filtering by Use of  Local  Statistics\",  Jong-Sen  Lee,\nIEEE Transactions on Pattern Analysis and Machine Intelligence, March 1980.\n\nThe edge enhancement details are from pgmenhance(1), which is taken from Philip R. Thompson's\n\"xim\" program, which in turn took it from section 6 of \"Digital Halftones by Dot  Diffusion\",\nD.  E.  Knuth,  ACM Transaction on Graphics Vol. 6, No. 4, October 1987, which in turn got it\nfrom two 1976 papers by J. F. Jarvis et. al.\n"
                }
            ]
        },
        "SEE ALSO": {
            "content": "pgmenhance(1), pnmconvol(1), pnm(5)\n",
            "subsections": []
        },
        "BUGS": {
            "content": "Integers and tables may overflow if PPMMAXMAXVAL is greater than 255.\n",
            "subsections": []
        },
        "AUTHOR": {
            "content": "Graeme W. Gill    graeme@labtam.oz.au\n\n\n\n5 February 1993                              pnmnlfilt(1)",
            "subsections": []
        }
    },
    "summary": "pnmnlfilt  -  non-linear filters: smooth, alpha trim mean, optimal estimation smoothing, edge enhancement.",
    "flags": [],
    "examples": [],
    "see_also": [
        {
            "name": "pgmenhance",
            "section": "1",
            "url": "https://www.chedong.com/phpMan.php/man/pgmenhance/1/json"
        },
        {
            "name": "pnmconvol",
            "section": "1",
            "url": "https://www.chedong.com/phpMan.php/man/pnmconvol/1/json"
        },
        {
            "name": "pnm",
            "section": "5",
            "url": "https://www.chedong.com/phpMan.php/man/pnm/5/json"
        }
    ],
    "tldr": {
        "source": "official",
        "description": "Apply a non-linear filter onto a PNM image.",
        "examples": [
            {
                "description": "Apply the \"alpha trimmed mean\" filter with the specified alpha and radius values onto the PNM image",
                "command": "pnmnlfilt {{0.0..0.5}} {{radius}} {{path/to/image.pnm}} > {{path/to/output.pnm}}"
            },
            {
                "description": "Apply the \"optimal estimation smoothing\" filter with the specified noise threshold and radius onto the PNM image",
                "command": "pnmnlfilt {{1.0..2.0}} {{radius}} {{path/to/image.pnm}} > {{path/to/output.pnm}}"
            },
            {
                "description": "Apply the \"edge enhancement\" filter with the specified alpha and radius onto the PNM image",
                "command": "pnmnlfilt {{0.9..(-0.1)}} {{radius}} {{path/to/image.pnm}} > {{path/to/output.pnm}}"
            }
        ]
    }
}