FILTM

Mapping to filter objects in datasets and datafiles

     [B,OUT] = FILTM(A,FILTER_COMMAND,{PAR1,PAR2,....},SIZE)
     [B,OUT] = A*FILTM([],FILTER_COMMAND,{PAR1,PAR2,....},SIZE)

Input
A	Dataset or datafile or double
FILTER_COMMAND	String with function name
{PAR1,	... } Cell array with optional parameters to FILTER_COMMAND
SIZE	Output size of the mapping (default: input size)

Output
B	Dataset or datafile of images processed by FILTER_COMMAND
OUT	Possible additional output, only if A is double

Description

For each object stored in A a filter operation is performed as

      OBJECT_OUT = FILTER_COMMAND(OBJECT_IN,PAR1,PAR2,....)

The results are collected and stored in B. In case A (and thereby B) is  a datafile, execution is postponed until conversion into a dataset, or a  call to SAVEDATAFILE or CREATEDATAFILE.

Just for fun: in case A is a double, more parameters may be returned.  See below for an example.

Example(s)

 b = filtm(a,'conv2',{[-1 0 1; -1 0 1; -1 0 1],'same'});
 Performs a convolution with a horizontal gradient filter (see CONV2).

 argmin2 = filtm([],'min',{[],2})*out2 % Define argmin for dimension 2
 a = rand(10,4)                        % some data
 a*argmin2                             % return indices of minima

 There is a similar command FILTIM that is recommended for handling
 multi-band images.

See also

datasets, datafiles, im2obj, data2im, im2feat, datgauss, datfilt, filtim, savedatafile, createdatafile,