pamlookup -lookupfile=lookupfile [-byplane] -missingcolor=color [-fit] indexfile
All options can be abbreviated to their shortest unique prefix. You may use two hyphens instead of one. You may separate an option name and its value with white space instead of an equals sign.
This program is part of Netpbm(1).
pamlookup takes a two dimensional array of indices and a lookup table as input. For each position in the index array, it looks up the index in the lookup table and places the result of the lookup in the output image.
There are two ways of indexing the lookup table: whole tuple and by plane. The -byplane option controls which pamlookup does.
In the simplest form of whole tuple indexing, each index in the index array is a single whole number and the lookup table associates a whole tuple with each possible whole number index. So for example, the index array might have at Row 2, Column 9 the value 23. The lookup table might associate the tuple (1,2,3) with the value 23. In that case, the output image contains the tuple (1,2,3) at Row 2, Column 9.
In a more complex form of whole tuple indexing, each index in the index array is an ordered pair of whole numbers and the lookup table associates a whole tuple with each possible ordered pair index. Modifying the example above, the index value could be (23, 5) instead of 23.
With whole tuple indexing, the output thus has the same width and height as the index image, and tuple depth and type and maxval determined by the lookup table.
With whole tuple indexing, if the index image has depth 1, its sample values are single whole number indices. If the index image has depth greater than 1, its tuples are ordered pair indices composed of the first and second sample in the tuple.
In by plane indexing, the index image contains whole number indices. The first sample of each tuple in the index image is the index. The lookup table maps each whole number index to another whole number. pamlookup looks up each sample from the index image in the lookup table and uses the resulting whole number as the sample value for the same row, column, and plane in the output.
With by plane indexing, the output thus has the same dimensions as the index image an the same maxval as the lookup image.
The Lookup Table Image
The lookup table is a PAM or PNM image. If the index image contains whole number indices, the lookup image is a single row and the index is a column number. The lookup result is the value of the tuple or pixel at the indicated column in the one row in the lookup table. If the index image contains ordered pair indices, the first element of the ordered pair is a row number and the second element of the ordered pair is a column number. The lookup result is the value of the tuple or pixel at the indicated row and column in the lookup table.
The width of the lookup image should normally be the maxval of the index image plus one, so that each possible index sample value corresponds to one entry in the lookup table. There are two ways pamlookup deals with a lookup image that does not have such a width:
- Scale the lookup image to the required width. pamlookup always does this with by plane indexing, and with whole tuple indexing, does it when you specify -fit.
Use a default value for indices that exceed the width of the lookup image and ignore lookup image columns beyond the maxval of the index image. pamlookup does this with whole tuple indexing when you don't specify -fit.
You specify the default value with a -missingcolor option; it defaults to the value from the top left corner of the lookup image.
With ordered pair indexes (which implies whole tuple indexing), the same rule applies to the height of the index image as to the width.
The mandatory -lookupfile option identifies the file containing the lookup table image. - means Standard Input. It won't work if both the index image file and lookup table file are Standard Input.
You can use ppmmake and pnmcat to create a lookup table file.
Example - Whole Tuple Indexing
Here is an example of pamlookup's function with whole tuple indexing (-byplane not specified).
Consider an index image consisting of a 3x2x1 PAM as follows:
and a lookup table consisting of a 3x1 PPM image as follows:
The lookup table above says Index 0 corresponds to the color red, Index 1 corresponds to yellow, and Index 2 corresponds to beige. The output of pamlookup is the following PPM image:
Now let's look at an example of the more complex case where the indices are ordered pairs of whole numbers instead of whole numbers. Our index image will be this 3x2x2 PAM image:
Our lookup table for the example will be this two dimensional PPM:
Example - By Plane Indexing
Here is an example of pamlookup's function with by plane tuple indexing (-byplane specified).
Consider an index image consisting of a 3x2x3 PAM as follows:
and a lookup table consisting of a 3x1x1 PAM image with maxval 7 as follows:
The lookup table above says Index 0 corresponds to the sample value 3, Index 1 corresponds to 4, and Index 2 corresponds to 7. The output of pamlookup is the following 3x2x3 PAM image:
The indexfile argument identifies the file containing the index PAM or PNM image. - means Standard Input. It won't work if both the index image file and lookup table file are Standard Input.
The output image goes to Standard Output.
If you want to use two separate 1-plane images as indices (so that your output reflects the combination of both inputs), use pamstack to combine the two into one two-plane image (and use a 2-dimensional lookup table image).
In addition to the options common to all programs based on libnetpbm (most notably -quiet, see Common Options ), pamlookup recognizes the following command line options:
lookupfile names the file that contains the PAM or PNM image that is the lookup table. This option is mandatory.
This options selects by plane indexing. The default is whole tuple indexing.
This option was new in Netpbm 10.72 (September 2015). Before that, there is no by plane indexing.
This option is meaningful only if the lookup image (and therefore the output) is a PNM image. color specifies the color that is to go in the output wherever the index from the input is not present in the lookup table (not present means the index exceeds the dimensions of the lookup image -- e.g. index is 100 but the lookup image is a 50 x 1 PPM).
If you don't specify this option or -fit, pamlookup uses the value from the top left corner of the lookup image whenever an index exceeds the dimensions of the lookup image.
Specify the color (color) as described for the argument of the pnm_parsecolor() library routine .
Another way to deal with a too-small lookup image is to use the -fit option.
This option has no effect if you also specify -fit or -byplane.
This option says to shrink or expand the lookup image as necessary to fit the indices present in the index image, per the index image's maxval. For example, if your index image has a single plane and a maxval of 255 and your lookup image is 1 row of 10 columns, pamlookup stretches your lookup image to 255 columns before doing the lookups. pamlookup does the stretching (or shrinking) with the pamscale(1) program.
When you use -fit, pamlookup never fails or warns you because of invalid lookup image dimensions, and the -missingcolor option has no effect.
-fit has no effect when you specify -byplane. pamlookup always has the behavior requested by -fit when it does by plane indexing.
Example: rainfall map
Say you have a set of rainfall data in a single plane PAM image. The rows and columns of the PAM indicate latitude and longitude. The sample values are the annual rainfall in (whole) centimeters. The highest rainfall value in the image is 199 centimeters. The image is in the file rainfall.pam.
You want to produce a PPM rainfall map with green for the wettest places, red for the driest, and other colors in between.
First, compose a lookup table image, probably with a graphical editor and the image blown way up so you can work with individual pixels. The image must have a single row and 200 columns. Make the leftmost pixel red and the rightmost pixel green and choose appropriate colors in between. Call it colorkey.ppm.
pamlookup rainfall.pam -lookupfile=colorkey.ppm >rainfallmap.ppm
Now lets say you're too lazy to type in 200 color values and nobody really cares about the places that have more than 99 centimeters of annual rainfall. In that case, just make colorkey.ppm 100 columns wide and do this:
pamlookup rainfall.ppm -lookupfile=colorkey.ppm -missingcolor=black \ >rainfallmap.ppm
Now if there are areas that get more than 100 centimeters of rainfall, they will just show up black in the output.
Example: graphical diff
Say you want to compare two PBM (black and white) images visually. Each consists of black foreground pixels on a white background. You want to create an image that contains background where both images contain background and foreground where both images contain foreground. But where Image 1 has a foreground pixel and Image 2 does not, you want red in the output; where Image 2 has a foreground pixel and Image 1 does not, you want green.
First, we create a single image that contains the information from both input PBMs:
pamstack image1.pbm image2.pbm >bothimages.pam
Note that this image has 1 of 4 possible tuple values at each location: (0,0), (0,1), (1,0), or (1,1).
Now, we create a lookup table that we can index with those 4 values:
ppmmake white 1 1 >white.ppm ppmmake black 1 1 >black.ppm ppmmake red 1 1 >red.ppm ppmmake green 1 1 >green.ppm pnmcat -leftright black.ppm red.ppm >blackred.ppm pnmcat -leftright green.ppm white.ppm >greenwhite.ppm pnmcat -topbottom blackred.ppm greenwhite.ppm >lookup.ppm
Finally, we look up the indices from our index in our lookup table and produce the output:
pamlookup bothimages.ppm -lookupfile=lookup.ppm >imagediff.ppm
pamunlookup(1), pnmremap(1), ppmmake(1), pnmcat(1), pamstack(1), pnm(1), pam(1)
pamlookup was new in Netpbm 10.13 (December 2002).
This manual page was generated by the Netpbm tool 'makeman' from HTML source. The master documentation is at
pamtris(1), pamunlookup(1), pnmremap(1).