Mercurial > hg > octave-image
annotate inst/fspecial.m @ 857:75d6748324de
New functions imgradient and imgradientxy (patch #8265)
* imgradient.m, imgradientxy.m: new function files.
* INDEX: update list of functions.
* NEWS: update list of new functinos for next release.
| author | Brandon Miles <brandon.miles7@gmail.com> |
|---|---|
| date | Sun, 05 Jan 2014 07:27:06 +0000 |
| parents | 71a92303203d |
| children | f1cf02800a87 |
| rev | line source |
|---|---|
| 676 | 1 ## Copyright (C) 2005 Søren Hauberg <soren@hauberg.org> |
|
561
c45838839d86
maint: update license to GPLv3 and mention non GPL files
carandraug
parents:
356
diff
changeset
|
2 ## |
|
c45838839d86
maint: update license to GPLv3 and mention non GPL files
carandraug
parents:
356
diff
changeset
|
3 ## This program is free software; you can redistribute it and/or modify it under |
|
c45838839d86
maint: update license to GPLv3 and mention non GPL files
carandraug
parents:
356
diff
changeset
|
4 ## the terms of the GNU General Public License as published by the Free Software |
|
c45838839d86
maint: update license to GPLv3 and mention non GPL files
carandraug
parents:
356
diff
changeset
|
5 ## Foundation; either version 3 of the License, or (at your option) any later |
|
c45838839d86
maint: update license to GPLv3 and mention non GPL files
carandraug
parents:
356
diff
changeset
|
6 ## version. |
|
c45838839d86
maint: update license to GPLv3 and mention non GPL files
carandraug
parents:
356
diff
changeset
|
7 ## |
|
c45838839d86
maint: update license to GPLv3 and mention non GPL files
carandraug
parents:
356
diff
changeset
|
8 ## This program is distributed in the hope that it will be useful, but WITHOUT |
|
c45838839d86
maint: update license to GPLv3 and mention non GPL files
carandraug
parents:
356
diff
changeset
|
9 ## ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or |
|
c45838839d86
maint: update license to GPLv3 and mention non GPL files
carandraug
parents:
356
diff
changeset
|
10 ## FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more |
|
c45838839d86
maint: update license to GPLv3 and mention non GPL files
carandraug
parents:
356
diff
changeset
|
11 ## details. |
|
c45838839d86
maint: update license to GPLv3 and mention non GPL files
carandraug
parents:
356
diff
changeset
|
12 ## |
|
c45838839d86
maint: update license to GPLv3 and mention non GPL files
carandraug
parents:
356
diff
changeset
|
13 ## You should have received a copy of the GNU General Public License along with |
|
c45838839d86
maint: update license to GPLv3 and mention non GPL files
carandraug
parents:
356
diff
changeset
|
14 ## this program; if not, see <http://www.gnu.org/licenses/>. |
| 217 | 15 |
|
220
df13bd973471
Documentation is now in Texinfo (looks better on the website)
hauberg
parents:
217
diff
changeset
|
16 ## -*- texinfo -*- |
|
df13bd973471
Documentation is now in Texinfo (looks better on the website)
hauberg
parents:
217
diff
changeset
|
17 ## @deftypefn {Function File} {@var{filter} = } fspecial(@var{type}, @var{arg1}, @var{arg2}) |
|
df13bd973471
Documentation is now in Texinfo (looks better on the website)
hauberg
parents:
217
diff
changeset
|
18 ## Create spatial filters for image processing. |
|
df13bd973471
Documentation is now in Texinfo (looks better on the website)
hauberg
parents:
217
diff
changeset
|
19 ## |
|
df13bd973471
Documentation is now in Texinfo (looks better on the website)
hauberg
parents:
217
diff
changeset
|
20 ## @var{type} determines the shape of the filter and can be |
| 356 | 21 ## @table @t |
| 22 ## @item "average" | |
| 23 ## Rectangular averaging filter. The optional argument @var{arg1} controls the | |
| 24 ## size of the filter. If @var{arg1} is an integer @var{N}, a @var{N} by @var{N} | |
| 25 ## filter is created. If it is a two-vector with elements @var{N} and @var{M}, the | |
| 26 ## resulting filter will be @var{N} by @var{M}. By default a 3 by 3 filter is | |
| 27 ## created. | |
| 28 ## @item "disk" | |
| 29 ## Circular averaging filter. The optional argument @var{arg1} controls the | |
| 30 ## radius of the filter. If @var{arg1} is an integer @var{N}, a 2 @var{N} + 1 | |
| 31 ## filter is created. By default a radius of 5 is used. | |
| 32 ## @item "gaussian" | |
| 33 ## Gaussian filter. The optional argument @var{arg1} controls the size of the | |
| 34 ## filter. If @var{arg1} is an integer @var{N}, a @var{N} by @var{N} | |
| 35 ## filter is created. If it is a two-vector with elements @var{N} and @var{M}, the | |
| 36 ## resulting filter will be @var{N} by @var{M}. By default a 3 by 3 filter is | |
| 37 ## created. The optional argument @var{arg2} sets spread of the filter. By default | |
| 38 ## a spread of @math{0.5} is used. | |
| 39 ## @item "log" | |
| 40 ## Laplacian of Gaussian. The optional argument @var{arg1} controls the size of the | |
| 41 ## filter. If @var{arg1} is an integer @var{N}, a @var{N} by @var{N} | |
| 42 ## filter is created. If it is a two-vector with elements @var{N} and @var{M}, the | |
| 43 ## resulting filter will be @var{N} by @var{M}. By default a 5 by 5 filter is | |
| 44 ## created. The optional argument @var{arg2} sets spread of the filter. By default | |
| 45 ## a spread of @math{0.5} is used. | |
| 46 ## @item "laplacian" | |
| 47 ## 3x3 approximation of the laplacian. The filter is approximated as | |
| 48 ## @example | |
| 49 ## (4/(@var{alpha}+1))*[@var{alpha}/4, (1-@var{alpha})/4, @var{alpha}/4; ... | |
| 50 ## (1-@var{alpha})/4, -1, (1-@var{alpha})/4; ... | |
| 51 ## @var{alpha}/4, (1-@var{alpha})/4, @var{alpha}/4]; | |
| 52 ## @end example | |
| 53 ## where @var{alpha} is a number between 0 and 1. This number can be controlled | |
| 54 ## via the optional input argument @var{arg1}. By default it is @math{0.2}. | |
| 55 ## @item "unsharp" | |
| 56 ## Sharpening filter. The following filter is returned | |
| 57 ## @example | |
| 58 ## (1/(@var{alpha}+1))*[-@var{alpha}, @var{alpha}-1, -@var{alpha}; ... | |
| 59 ## @var{alpha}-1, @var{alpha}+5, @var{alpha}-1; ... | |
| 60 ## -@var{alpha}, @var{alpha}-1, -@var{alpha}]; | |
| 61 ## @end example | |
| 62 ## where @var{alpha} is a number between 0 and 1. This number can be controlled | |
| 63 ## via the optional input argument @var{arg1}. By default it is @math{0.2}. | |
| 64 ## @item "motion" | |
| 65 ## Moion blur filter of width 1 pixel. The optional input argument @var{arg1} | |
| 66 ## controls the length of the filter, which by default is 9. The argument @var{arg2} | |
| 67 ## controls the angle of the filter, which by default is 0 degrees. | |
| 68 ## @item "sobel" | |
| 69 ## Horizontal Sobel edge filter. The following filter is returned | |
| 70 ## @example | |
| 71 ## [ 1, 2, 1; | |
| 72 ## 0, 0, 0; | |
| 73 ## -1, -2, -1 ] | |
| 74 ## @end example | |
| 75 ## @item "prewitt" | |
| 76 ## Horizontal Prewitt edge filter. The following filter is returned | |
| 77 ## @example | |
| 78 ## [ 1, 1, 1; | |
| 79 ## 0, 0, 0; | |
| 80 ## -1, -1, -1 ] | |
| 81 ## @end example | |
| 82 ## @item "kirsch" | |
| 83 ## Horizontal Kirsch edge filter. The following filter is returned | |
| 84 ## @example | |
| 85 ## [ 3, 3, 3; | |
| 86 ## 3, 0, 3; | |
| 87 ## -5, -5, -5 ] | |
| 88 ## @end example | |
|
220
df13bd973471
Documentation is now in Texinfo (looks better on the website)
hauberg
parents:
217
diff
changeset
|
89 ## @end table |
|
df13bd973471
Documentation is now in Texinfo (looks better on the website)
hauberg
parents:
217
diff
changeset
|
90 ## @end deftypefn |
|
df13bd973471
Documentation is now in Texinfo (looks better on the website)
hauberg
parents:
217
diff
changeset
|
91 |
| 676 | 92 ## Remarks by Søren Hauberg (jan. 2nd 2007) |
|
220
df13bd973471
Documentation is now in Texinfo (looks better on the website)
hauberg
parents:
217
diff
changeset
|
93 ## The motion filter and most of the documentation was taken from Peter Kovesi's |
|
df13bd973471
Documentation is now in Texinfo (looks better on the website)
hauberg
parents:
217
diff
changeset
|
94 ## GPL'ed implementation of fspecial from |
|
df13bd973471
Documentation is now in Texinfo (looks better on the website)
hauberg
parents:
217
diff
changeset
|
95 ## http://www.csse.uwa.edu.au/~pk/research/matlabfns/OctaveCode/fspecial.m |
|
df13bd973471
Documentation is now in Texinfo (looks better on the website)
hauberg
parents:
217
diff
changeset
|
96 |
| 356 | 97 function f = fspecial (type, arg1, arg2) |
| 98 if (!ischar (type)) | |
| 99 error ("fspecial: first argument must be a string"); | |
| 217 | 100 endif |
| 101 | |
| 102 switch lower(type) | |
| 103 case "average" | |
| 104 ## Get filtersize | |
| 356 | 105 if (nargin > 1 && isreal (arg1) && length (arg1 (:)) <= 2) |
| 106 fsize = arg1 (:); | |
| 217 | 107 else |
| 108 fsize = 3; | |
| 109 endif | |
| 110 ## Create the filter | |
| 356 | 111 f = ones (fsize); |
| 217 | 112 ## Normalize the filter to integral 1 |
| 356 | 113 f = f / sum (f (:)); |
| 217 | 114 |
| 115 case "disk" | |
| 116 ## Get the radius | |
| 356 | 117 if (nargin > 1 && isreal (arg1) && length (arg1 (:)) == 1) |
| 217 | 118 radius = arg1; |
| 119 else | |
| 120 radius = 5; | |
| 121 endif | |
| 122 ## Create the filter | |
| 356 | 123 [x, y] = meshgrid (-radius:radius, -radius:radius); |
| 124 r = sqrt (x.^2 + y.^2); | |
| 217 | 125 f = (r <= radius); |
| 126 ## Normalize the filter to integral 1 | |
| 356 | 127 f = f / sum (f (:)); |
| 217 | 128 |
| 129 case "gaussian" | |
| 130 ## Get hsize | |
| 356 | 131 if (nargin > 1 && isreal (arg1)) |
| 132 if (length (arg1 (:)) == 1) | |
| 217 | 133 hsize = [arg1, arg1]; |
| 356 | 134 elseif (length (arg1 (:)) == 2) |
| 217 | 135 hsize = arg1; |
| 136 else | |
| 356 | 137 error ("fspecial: second argument must be a scalar or a vector of two scalars"); |
| 217 | 138 endif |
| 139 else | |
| 140 hsize = [3, 3]; | |
| 141 endif | |
| 142 ## Get sigma | |
| 356 | 143 if (nargin > 2 && isreal (arg2) && length (arg2 (:)) == 1) |
| 217 | 144 sigma = arg2; |
| 145 else | |
| 146 sigma = 0.5; | |
| 147 endif | |
| 356 | 148 h1 = hsize (1)-1; h2 = hsize (2)-1; |
| 217 | 149 [x, y] = meshgrid(0:h2, 0:h1); |
| 150 x = x-h2/2; y = y-h1/2; | |
| 151 gauss = exp( -( x.^2 + y.^2 ) / (2*sigma^2) ); | |
| 356 | 152 f = gauss / sum (gauss (:)); |
| 217 | 153 |
| 154 case "laplacian" | |
| 155 ## Get alpha | |
| 356 | 156 if (nargin > 1 && isscalar (arg1)) |
| 217 | 157 alpha = arg1; |
| 356 | 158 if (alpha < 0 || alpha > 1) |
| 159 error ("fspecial: second argument must be between 0 and 1"); | |
| 217 | 160 endif |
| 161 else | |
| 162 alpha = 0.2; | |
| 163 endif | |
| 164 ## Compute filter | |
| 356 | 165 f = (4/(alpha+1))*[alpha/4, (1-alpha)/4, alpha/4; ... |
| 166 (1-alpha)/4, -1, (1-alpha)/4; ... | |
| 167 alpha/4, (1-alpha)/4, alpha/4]; | |
| 217 | 168 case "log" |
| 169 ## Get hsize | |
| 356 | 170 if (nargin > 1 && isreal (arg1)) |
| 171 if (length (arg1 (:)) == 1) | |
| 217 | 172 hsize = [arg1, arg1]; |
| 356 | 173 elseif (length (arg1 (:)) == 2) |
| 217 | 174 hsize = arg1; |
| 175 else | |
| 356 | 176 error ("fspecial: second argument must be a scalar or a vector of two scalars"); |
| 217 | 177 endif |
| 178 else | |
| 179 hsize = [5, 5]; | |
| 180 endif | |
| 181 ## Get sigma | |
| 356 | 182 if (nargin > 2 && isreal (arg2) && length (arg2 (:)) == 1) |
| 217 | 183 sigma = arg2; |
| 184 else | |
| 185 sigma = 0.5; | |
| 186 endif | |
| 187 ## Compute the filter | |
| 356 | 188 h1 = hsize (1)-1; h2 = hsize (2)-1; |
| 189 [x, y] = meshgrid(0:h2, 0:h1); | |
| 217 | 190 x = x-h2/2; y = y = y-h1/2; |
| 191 gauss = exp( -( x.^2 + y.^2 ) / (2*sigma^2) ); | |
| 192 f = ( (x.^2 + y.^2 - 2*sigma^2).*gauss )/( 2*pi*sigma^6*sum(gauss(:)) ); | |
| 193 | |
| 194 case "motion" | |
| 195 ## Taken (with some changes) from Peter Kovesis implementation | |
| 196 ## (http://www.csse.uwa.edu.au/~pk/research/matlabfns/OctaveCode/fspecial.m) | |
| 197 ## FIXME: The implementation is not quite matlab compatible. | |
| 356 | 198 if (nargin > 1 && isreal (arg1)) |
| 217 | 199 len = arg1; |
| 200 else | |
| 201 len = 9; | |
| 202 endif | |
| 356 | 203 if (mod (len, 2) == 1) |
| 217 | 204 sze = [len, len]; |
| 205 else | |
| 206 sze = [len+1, len+1]; | |
| 207 end | |
| 356 | 208 if (nargin > 2 && isreal (arg2)) |
| 217 | 209 angle = arg2; |
| 210 else | |
| 211 angle = 0; | |
| 212 endif | |
| 600 | 213 |
| 217 | 214 ## First generate a horizontal line across the middle |
| 356 | 215 f = zeros (sze); |
| 216 f (floor (len/2)+1, 1:len) = 1; | |
| 217 | 217 |
| 218 # Then rotate to specified angle | |
| 356 | 219 f = imrotate (f, angle, "bilinear", "loose"); |
| 220 f = f / sum (f (:)); | |
| 217 | 221 |
| 222 case "prewitt" | |
| 223 ## The filter | |
| 356 | 224 f = [1, 1, 1; 0, 0, 0; -1, -1, -1]; |
| 217 | 225 |
| 226 case "sobel" | |
| 227 ## The filter | |
| 356 | 228 f = [1, 2, 1; 0, 0, 0; -1, -2, -1]; |
| 217 | 229 |
| 230 case "kirsch" | |
| 231 ## The filter | |
| 356 | 232 f = [3, 3, 3; 3, 0, 3; -5, -5, -5]; |
| 217 | 233 |
| 234 case "unsharp" | |
| 235 ## Get alpha | |
| 356 | 236 if (nargin > 1 && isscalar (arg1)) |
| 217 | 237 alpha = arg1; |
| 356 | 238 if (alpha < 0 || alpha > 1) |
| 239 error ("fspecial: second argument must be between 0 and 1"); | |
| 217 | 240 endif |
| 241 else | |
| 242 alpha = 0.2; | |
| 243 endif | |
| 244 ## Compute filter | |
| 356 | 245 f = (1/(alpha+1))*[-alpha, alpha-1, -alpha; ... |
| 246 alpha-1, alpha+5, alpha-1; ... | |
| 247 -alpha, alpha-1, -alpha]; | |
| 217 | 248 |
| 249 otherwise | |
| 356 | 250 error ("fspecial: filter type '%s' is not supported", type); |
| 217 | 251 endswitch |
| 252 endfunction |