Mercurial > hg > octave-terminal
changeset 14209:846273dae16b stable
Return correct part of convolution for 'same' parameter in conv2, convn (Bug #34893).
* oct-convn.cc: Fix off-by-1 index error for 'same' part of convolution.
* conv2.cc (conv2, convn): Update documentation strings to be explicit about
what part of the convolution is returned for each SHAPE parameter. Add
new tests for conv2().
author | Rik <octave@nomad.inbox5.com> |
---|---|
date | Tue, 17 Jan 2012 14:37:45 -0800 |
parents | fe2ace7a35da |
children | a0d98842a4d8 1facbe04b00c |
files | liboctave/oct-convn.cc src/DLD-FUNCTIONS/conv2.cc |
diffstat | 2 files changed, 73 insertions(+), 27 deletions(-) [+] |
line wrap: on
line diff
--- a/liboctave/oct-convn.cc +++ b/liboctave/oct-convn.cc @@ -137,7 +137,7 @@ Array<idx_vector> sidx (dim_vector (nd, 1)); for (int i = 0; i < nd; i++) - sidx(i) = idx_vector::make_range ((bdims(i)-1)/2, 1, adims(i)); + sidx(i) = idx_vector::make_range (bdims(i)/2, 1, adims(i)); c = c.index (sidx); }
--- a/src/DLD-FUNCTIONS/conv2.cc +++ b/src/DLD-FUNCTIONS/conv2.cc @@ -34,24 +34,6 @@ enum Shape { SHAPE_FULL, SHAPE_SAME, SHAPE_VALID }; -/* -%!test -%! b = [0,1,2,3;1,8,12,12;4,20,24,21;7,22,25,18]; -%! assert(conv2([0,1;1,2],[1,2,3;4,5,6;7,8,9]),b); - -%!test -%! b = single([0,1,2,3;1,8,12,12;4,20,24,21;7,22,25,18]); -%! assert(conv2(single([0,1;1,2]),single([1,2,3;4,5,6;7,8,9])),b); - -%!assert (conv2 (1:3, 1:2, [1,2;3,4;5,6]), -%! [1,4,4;5,18,16;14,48,40;19,62,48;15,48,36;]); - -%!assert (conv2 (1:3, 1:2, [1,2;3,4;5,6], "full"), -%! conv2 (1:3, 1:2, [1,2;3,4;5,6])); - -*/ - - DEFUN_DLD (conv2, args, , "-*- texinfo -*-\n\ @deftypefn {Loadable Function} {} conv2 (@var{A}, @var{B})\n\ @@ -67,21 +49,24 @@ \n\ @item @var{shape} = \"same\"\n\ Return the central part of the convolution with the same size as @var{A}.\n\ +The central part of the convolution begins at the indices\n\ +@code{floor ([size(@var{B})/2] + 1)}.\n\ \n\ @item @var{shape} = \"valid\"\n\ Return only the parts which do not include zero-padded edges.\n\ +The size of the result is @code{max (size (A) - size (B) + 1, 0)}.\n\ @end table\n\ \n\ When the third argument is a matrix, return the convolution of the matrix\n\ @var{m} by the vector @var{v1} in the column direction and by the vector\n\ -@var{v2} in the row direction\n\ +@var{v2} in the row direction.\n\ @seealso{conv, convn}\n\ @end deftypefn") { octave_value retval; octave_value tmp; int nargin = args.length (); - std::string shape = "full"; //default + std::string shape = "full"; // default bool separable = false; convn_type ct; @@ -237,31 +222,88 @@ return retval; } +/* +%!test +%! c = [0,1,2,3;1,8,12,12;4,20,24,21;7,22,25,18]; +%! assert (conv2 ([0,1;1,2], [1,2,3;4,5,6;7,8,9]), c); + +%!test +%! c = single ([0,1,2,3;1,8,12,12;4,20,24,21;7,22,25,18]); +%! assert (conv2 (single ([0,1;1,2]), single ([1,2,3;4,5,6;7,8,9])), c); + +%!test +%! c = [1,4,4;5,18,16;14,48,40;19,62,48;15,48,36]; +%! assert (conv2 (1:3, 1:2, [1,2;3,4;5,6]), c); + +%!assert (conv2 (1:3, 1:2, [1,2;3,4;5,6], "full"), +%! conv2 (1:3, 1:2, [1,2;3,4;5,6])); + +%% Test shapes +%!shared A, B, C +%! A = rand (3, 4); +%! B = rand (4); +%! C = conv2 (A, B); +%!assert (conv2 (A,B, "full"), C) +%!assert (conv2 (A,B, "same"), C(3:5,3:6)) +%!assert (conv2 (A,B, "valid"), zeros (0, 1)) +%!assert (size (conv2 (B,A, "valid")), [2 1]) + +%!test +%! B = rand (5); +%! C = conv2 (A, B); +%!assert (conv2 (A,B, "full"), C) +%!assert (conv2 (A,B, "same"), C(3:5,3:6)) +%!assert (conv2 (A,B, "valid"), zeros (0, 0)) +%!assert (size (conv2 (B,A, "valid")), [3 2]) + +%% Clear shared variables so they are not reported for tests below +%!shared + +%% Test cases from Bug #34893 +%!assert (conv2 ([1:5;1:5], [1:2], 'same'), [4 7 10 13 10; 4 7 10 13 10]) +%!assert (conv2 ([1:5;1:5]', [1:2]', 'same'), [4 7 10 13 10; 4 7 10 13 10]') +%!#assert (conv2 ([1:5;1:5], [1:2], 'valid'), [4 7 10 13; 4 7 10 13]) +%!assert (conv2 ([1:5;1:5]', [1:2]', 'valid'), [4 7 10 13; 4 7 10 13]') + +%% Test input validation +%!error conv2 () +%!error conv2 (1) +%!error <SHAPE type not valid> conv2 (1,2, "NOT_A_SHAPE") +%% Test alternate calling form which should be 2 vectors and a matrix +%!error conv2 (ones (2), 1, 1) +%!error conv2 (1, ones (2), 1) + +*/ + DEFUN_DLD (convn, args, , "-*- texinfo -*-\n\ -@deftypefn {Loadable Function} {@var{C} =} convn (@var{A}, @var{B}, @var{shape})\n\ -Return the n-D convolution of @var{A} and @var{B} where the size\n\ -of @var{C} is given by\n\ +@deftypefn {Loadable Function} {@var{C} =} convn (@var{A}, @var{B})\n\ +@deftypefnx {Loadable Function} {@var{C} =} convn (@var{A}, @var{B}, @var{shape})\n\ +Return the n-D convolution of @var{A} and @var{B}. The size of the result\n\ +is determined by the optional @var{shape} argument which takes the following\n\ +values\n\ \n\ @table @asis\n\ @item @var{shape} = \"full\"\n\ -Return the full convolution.\n\ +Return the full convolution. (default)\n\ \n\ @item @var{shape} = \"same\"\n\ Return central part of the convolution with the same size as @var{A}.\n\ +The central part of the convolution begins at the indices\n\ +@code{floor ([size(@var{B})/2] + 1)}.\n\ \n\ @item @var{shape} = \"valid\"\n\ Return only the parts which do not include zero-padded edges.\n\ +The size of the result is @code{max (size (A) - size (B) + 1, 0)}.\n\ @end table\n\ \n\ -By default @var{shape} is @samp{\"full\"}.\n\ @seealso{conv2, conv}\n\ @end deftypefn") { octave_value retval; octave_value tmp; int nargin = args.length (); - std::string shape = "full"; //default + std::string shape = "full"; // default convn_type ct; if (nargin < 2 || nargin > 3) @@ -337,3 +379,7 @@ return retval; } + +/* + FIXME: Need tests for convn in addition to conv2. +*/