sergiotarxz
/
libvips
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

redo correlation funcs as classes 

im_fastcor()
im_spcor()
im_gradcor()

though gradcor we just deprecate, it's complex and hardly used ...
rework later is anyone complains
This commit is contained in:
John Cupitt 2013-11-08 14:09:42 +00:00
parent 9a06b2cea7
commit f0d4760560
28 changed files with 991 additions and 586 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
ChangeLog
View File

@ -1,7 +1,8 @@
19/10/13 started 7.37.0
- redone im_rotate_*mask45(), im_gauss_*mask*(), im_log_*mask(), im_dilate(),
im_erode(), im_rank_image(), im_compass(), im_linedet(), im_gradient(),
im_convsep(), im_convsep_f() as classes
im_convsep(), im_convsep_f(), im_fastcor(), im_spcor() as classes
- im_gradcor() deprecated
- vips_init() now does some ABI compat checking, though this change requires
an ABI break
- add "interlace" option to vips_jpegsave()

12
TODO
View File

@ -1,3 +1,15 @@
- look again at gcc auto-vectorisation, what would we need to do to use this?
- how about
top -d 0.1 | grep vips > log
to watch RES at 0.1s intervals?
- add "crop" as a synonym for "extract_area"
- add vips_gamma(), see rad etc.
- drag display control slider, broken
- do conv and morph quickly as simple wrappers over the vips7 operations

2
libvips/arithmetic/add.c
View File

@ -249,7 +249,7 @@ vips_add_init( VipsAdd *add )
* one-band image by joining n copies of the one-band image together, and then
* the two n-band images are operated upon.
*
* The two input images are cast up to the smallest common type (see table
* The two input images are cast up to the smallest common format (see table
* Smallest common format in
* <link linkend="VIPS-arithmetic">arithmetic</link>), then the
* following table is used to determine the output type:

2
libvips/arithmetic/boolean.c
View File

@ -258,7 +258,7 @@ vips_booleanv( VipsImage *left, VipsImage *right, VipsImage **out,
* one-band image by joining n copies of the one-band image together, and then
* the two n-band images are operated upon.
*
* The two input images are cast up to the smallest common type (see table
* The two input images are cast up to the smallest common format (see table
* Smallest common format in
* <link linkend="VIPS-arithmetic">arithmetic</link>).
*

2
libvips/arithmetic/divide.c
View File

@ -256,7 +256,7 @@ vips_divide_init( VipsDivide *divide )
* one-band image by joining n copies of the one-band image together, and then
* the two n-band images are operated upon.
*
* The two input images are cast up to the smallest common type (see table
* The two input images are cast up to the smallest common format (see table
* Smallest common format in
* <link linkend="VIPS-arithmetic">arithmetic</link>), then the
* following table is used to determine the output type:

2
libvips/arithmetic/linear.c
View File

@ -180,7 +180,7 @@ vips_linear_build( VipsObject *object )
for( x = 0; x < width; x++ ) \
for( k = 0; k < nb; k++ ) { \
q[0] = a[k] * p[0] + b[k]; \
q[1] = a[k] * p[1]; \
q[1] = p[1]; \
q += 2; \
p += 2; \
} \

2
libvips/arithmetic/math2.c
View File

@ -248,7 +248,7 @@ vips_math2v( VipsImage *left, VipsImage *right, VipsImage **out,
* one-band image by joining n copies of the one-band image together, and then
* the two n-band images are operated upon.
*
* The two input images are cast up to the smallest common type (see table
* The two input images are cast up to the smallest common format (see table
* Smallest common format in
* <link linkend="VIPS-arithmetic">arithmetic</link>), and that format is the
* result type.

6
libvips/arithmetic/multiply.c
View File

@ -158,7 +158,7 @@ vips_multiply_buffer( VipsArithmetic *arithmetic,
/* Type promotion for multiplication. Sign and value preserving. Make sure
* these match the case statement in multiply_buffer() above.
*/
static int vips_bandfmt_multiply[10] = {
static int vips_multiply_format_table[10] = {
/* UC C US S UI I F X D DX */
US, S, UI, I, UI, I, F, X, D, DX
};
@ -172,7 +172,7 @@ vips_multiply_class_init( VipsMultiplyClass *class )
object_class->nickname = "multiply";
object_class->description = _( "multiply two images" );
vips_arithmetic_set_format_table( aclass, vips_bandfmt_multiply );
vips_arithmetic_set_format_table( aclass, vips_multiply_format_table );
aclass->process_line = vips_multiply_buffer;
}
@ -198,7 +198,7 @@ vips_multiply_init( VipsMultiply *multiply )
* one-band image by joining n copies of the one-band image together, and then
* the two n-band images are operated upon.
*
* The two input images are cast up to the smallest common type (see table
* The two input images are cast up to the smallest common format (see table
* Smallest common format in
* <link linkend="VIPS-arithmetic">arithmetic</link>), then the
* following table is used to determine the output type:

2
libvips/arithmetic/relational.c
View File

@ -268,7 +268,7 @@ vips_relationalv( VipsImage *left, VipsImage *right, VipsImage **out,
* one-band image by joining n copies of the one-band image together, and then
* the two n-band images are operated upon.
*
* The two input images are cast up to the smallest common type (see table
* The two input images are cast up to the smallest common format (see table
* Smallest common format in
* <link linkend="VIPS-arithmetic">arithmetic</link>).
*

2
libvips/arithmetic/remainder.c
View File

@ -210,7 +210,7 @@ vips_remainder_init( VipsRemainder *remainder )
* one-band image by joining n copies of the one-band image together, and then
* the two n-band images are operated upon.
*
* The two input images are cast up to the smallest common type (see table
* The two input images are cast up to the smallest common format (see table
* Smallest common format in
* <link linkend="VIPS-arithmetic">arithmetic</link>), and that format is the
* result type.

2
libvips/arithmetic/subtract.c
View File

@ -188,7 +188,7 @@ vips_subtract_init( VipsSubtract *subtract )
* one-band image by joining n copies of the one-band image together, and then
* the two n-band images are operated upon.
*
* The two input images are cast up to the smallest common type (see table
* The two input images are cast up to the smallest common format (see table
* Smallest common format in
* <link linkend="VIPS-arithmetic">arithmetic</link>), then the
* following table is used to determine the output type:

2
libvips/conversion/embed.c
View File

@ -615,7 +615,7 @@ vips_embed_init( VipsEmbed *embed )
*
* Optional arguments:
*
* @extend: how to generate the edge pixels
* @extend: how to generate the edge pixels (default: black)
* @background: colour for edge pixels
*
* The opposite of vips_extract_area(): embed @in within an image of size

10
libvips/convolution/Makefile.am
View File

@ -3,18 +3,18 @@ noinst_LTLIBRARIES = libconvolution.la
libconvolution_la_SOURCES = \
convolution.c \
pconvolution.h \
correlation.c \
correlation.h \
conv.c \
convsep.c \
compass.c \
morph.c \
convol_dispatch.c \
fastcor.c \
spcor.c \
im_aconv.c \
im_aconvsep.c \
im_conv.c \
im_conv_f.c \
im_fastcor.c \
im_gradcor.c \
im_sharpen.c \
im_spcor.c
im_sharpen.c
AM_CPPFLAGS = -I${top_srcdir}/libvips/include @VIPS_CFLAGS@ @VIPS_INCLUDES@

4
libvips/convolution/convolution.c
View File

@ -150,9 +150,13 @@ vips_convolution_operation_init( void )
extern int vips_morph_get_type( void );
extern int vips_compass_get_type( void );
extern int vips_convsep_get_type( void );
extern int vips_fastcor_get_type( void );
extern int vips_spcor_get_type( void );
vips_conv_get_type();
vips_morph_get_type();
vips_compass_get_type();
vips_convsep_get_type();
vips_fastcor_get_type();
vips_spcor_get_type();
}

165
libvips/convolution/correlation.c Normal file
View File

@ -0,0 +1,165 @@
/* base class for correlation
*
* 7/11/13
* - from convolution.c
*/
/*
This file is part of VIPS.
VIPS is free software; you can redistribute it and/or modify
it under the terms of the GNU Lesser General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
02110-1301 USA
*/
/*
These files are distributed with VIPS - http://www.vips.ecs.soton.ac.uk
*/
#ifdef HAVE_CONFIG_H
#include <config.h>
#endif /*HAVE_CONFIG_H*/
#include <vips/intl.h>
#include <stdio.h>
#include <math.h>
#include <vips/vips.h>
#include <vips/internal.h>
#include "pconvolution.h"
#include "correlation.h"
G_DEFINE_ABSTRACT_TYPE( VipsCorrelation, vips_correlation,
VIPS_TYPE_OPERATION );
static int
vips_correlation_gen( VipsRegion *or,
void *seq, void *a, void *b, gboolean *stop )
{
VipsRegion *ir = (VipsRegion *) seq;
VipsCorrelation *correlation = (VipsCorrelation *) b;
VipsCorrelationClass *cclass =
VIPS_CORRELATION_GET_CLASS( correlation );
VipsRect *r = &or->valid;
VipsRect irect;
/* What part of ir do we need?
*/
irect.left = r->left;
irect.top = r->top;
irect.width = r->width + correlation->ref_ready->Xsize - 1;
irect.height = r->height + correlation->ref_ready->Ysize - 1;
if( vips_region_prepare( ir, &irect ) )
return( -1 );
cclass->correlation( correlation, ir, or );
return( 0 );
}
static int
vips_correlation_build( VipsObject *object )
{
VipsObjectClass *class = VIPS_OBJECT_GET_CLASS( object );
VipsCorrelationClass *cclass = VIPS_CORRELATION_CLASS( class );
VipsCorrelation *correlation = (VipsCorrelation *) object;
VipsImage **t = (VipsImage **) vips_object_local_array( object, 5 );
if( VIPS_OBJECT_CLASS( vips_correlation_parent_class )->
build( object ) )
return( -1 );
/* Stretch input out.
*/
if( vips_embed( correlation->in, &t[0],
correlation->ref->Xsize / 2, correlation->ref->Ysize / 2,
correlation->in->Xsize + correlation->ref->Xsize - 1,
correlation->in->Ysize + correlation->ref->Ysize - 1,
"extend", VIPS_EXTEND_COPY,
NULL ) ||
vips__formatalike( t[0], correlation->ref, &t[1], &t[2] ) ||
vips__bandalike( class->nickname, t[1], t[2], &t[3], &t[4] ) ||
vips_image_wio_input( t[4] ) )
return( -1 );
correlation->in_ready = t[3];
correlation->ref_ready = t[4];
g_object_set( object, "out", vips_image_new(), NULL );
/* FATSTRIP is good for us as THINSTRIP will cause
* too many recalculations on overlaps.
*/
if( vips_image_pipelinev( correlation->out,
VIPS_DEMAND_STYLE_FATSTRIP,
correlation->in_ready, correlation->ref_ready, NULL ) )
return( -1 );
correlation->out->Xsize = correlation->in->Xsize;
correlation->out->Ysize = correlation->in->Ysize;
correlation->out->BandFmt =
cclass->format_table[correlation->in_ready->BandFmt];
if( cclass->pre_generate &&
cclass->pre_generate( correlation ) )
return( -1 );
if( vips_image_generate( correlation->out,
vips_start_one, vips_correlation_gen, vips_stop_one,
correlation->in_ready, correlation ) )
return( -1 );
return( 0 );
}
static void
vips_correlation_class_init( VipsCorrelationClass *class )
{
GObjectClass *gobject_class = G_OBJECT_CLASS( class );
VipsObjectClass *object_class = (VipsObjectClass *) class;
gobject_class->set_property = vips_object_set_property;
gobject_class->get_property = vips_object_get_property;
object_class->nickname = "correlation";
object_class->description = _( "correlation operation" );
object_class->build = vips_correlation_build;
VIPS_ARG_IMAGE( class, "in", 0,
_( "Input" ),
_( "Input image argument" ),
VIPS_ARGUMENT_REQUIRED_INPUT,
G_STRUCT_OFFSET( VipsCorrelation, in ) );
VIPS_ARG_IMAGE( class, "ref", 10,
_( "Mask" ),
_( "Input reference image" ),
VIPS_ARGUMENT_REQUIRED_INPUT,
G_STRUCT_OFFSET( VipsCorrelation, ref ) );
VIPS_ARG_IMAGE( class, "out", 20,
_( "Output" ),
_( "Output image" ),
VIPS_ARGUMENT_REQUIRED_OUTPUT,
G_STRUCT_OFFSET( VipsCorrelation, out ) );
}
static void
vips_correlation_init( VipsCorrelation *correlation )
{
}

96
libvips/convolution/correlation.h Normal file
View File

@ -0,0 +1,96 @@
/* base class for all correlation operations
*/
/*
Copyright (C) 1991-2005 The National Gallery
This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
This library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with this library; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
02110-1301 USA
*/
/*
These files are distributed with VIPS - http://www.vips.ecs.soton.ac.uk
*/
#ifndef VIPS_PCORRELATION_H
#define VIPS_PCORRELATION_H
#ifdef __cplusplus
extern "C" {
#endif /*__cplusplus*/
#include <vips/vector.h>
#define VIPS_TYPE_CORRELATION (vips_correlation_get_type())
#define VIPS_CORRELATION( obj ) \
(G_TYPE_CHECK_INSTANCE_CAST( (obj), \
VIPS_TYPE_CORRELATION, VipsCorrelation ))
#define VIPS_CORRELATION_CLASS( klass ) \
(G_TYPE_CHECK_CLASS_CAST( (klass), \
VIPS_TYPE_CORRELATION, VipsCorrelationClass))
#define VIPS_IS_CORRELATION( obj ) \
(G_TYPE_CHECK_INSTANCE_TYPE( (obj), VIPS_TYPE_CORRELATION ))
#define VIPS_IS_CORRELATION_CLASS( klass ) \
(G_TYPE_CHECK_CLASS_TYPE( (klass), VIPS_TYPE_CORRELATION ))
#define VIPS_CORRELATION_GET_CLASS( obj ) \
(G_TYPE_INSTANCE_GET_CLASS( (obj), \
VIPS_TYPE_CORRELATION, VipsCorrelationClass ))
typedef struct {
VipsOperation parent_instance;
/* Params.
*/
VipsImage *in;
VipsImage *ref;
VipsImage *out;
/* The two input images, upcast to the smallest common format. ref is
* a memory buffer.
*/
VipsImage *in_ready;
VipsImage *ref_ready;
} VipsCorrelation;
typedef struct {
VipsOperationClass parent_class;
/* For each upcast input format, what output format.
*/
const VipsBandFormat *format_table;
/* Run just before generate. The subclass can fill in some stuff.
*/
int (*pre_generate)( VipsCorrelation * );
void (*correlation)( VipsCorrelation *,
VipsRegion *in, VipsRegion *out );
} VipsCorrelationClass;
GType vips_correlation_get_type( void );
#ifdef __cplusplus
}
#endif /*__cplusplus*/
#endif /*VIPS_PCORRELATION_H*/

267
libvips/convolution/fastcor.c Normal file
View File

@ -0,0 +1,267 @@
/* fastcor
*
* Copyright: 1990, N. Dessipris.
*
* Author: Nicos Dessipris
* Written on: 02/05/1990
* Modified on : 15/03/1991
* 20/2/95 JC
* - ANSIfied
* - in1 and in2 swapped, to match order for im_spcor
* - memory leaks fixed
* 21/2/95 JC
* - partialed
* - speed-ups
* 7/4/04
* - now uses im_embed() with edge stretching on the output
* - sets Xoffset / Yoffset
* 8/3/06 JC
* - use im_embed() with edge stretching on the input, not the output
* - calculate sum of squares of differences, rather than abs of
* difference
* 3/2/10
* - gtkdoc
* - cleanups
* 7/11/13
* - redone as a class
*/
/*
This file is part of VIPS.
VIPS is free software; you can redistribute it and/or modify
it under the terms of the GNU Lesser General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
02110-1301 USA
*/
/*
These files are distributed with VIPS - http://www.vips.ecs.soton.ac.uk
*/
#ifdef HAVE_CONFIG_H
#include <config.h>
#endif /*HAVE_CONFIG_H*/
#include <vips/intl.h>
#include <stdio.h>
#include <math.h>
#include <vips/vips.h>
#include "pconvolution.h"
#include "correlation.h"
typedef VipsCorrelationClass VipsFastcor;
typedef VipsCorrelationClass VipsFastcorClass;
G_DEFINE_TYPE( VipsFastcor, vips_fastcor, VIPS_TYPE_CORRELATION );
#define CORR_INT( TYPE ) { \
for( y = 0; y < r->height; y++ ) { \
unsigned int *q = (unsigned int *) \
VIPS_REGION_ADDR( out, r->left, r->top + y ); \
\
for( x = 0; x < r->width; x++ ) \
for( b = 0; b < bands; b++ ) { \
TYPE *p1 = (TYPE *) ref->data; \
TYPE *p2 = (TYPE *) VIPS_REGION_ADDR( in, \
r->left + x, r->top + y ); \
\
unsigned int sum; \
\
sum = 0; \
for( j = 0; j < ref->Ysize; j++ ) { \
for( i = b; i < sz; i += bands ) { \
int t = p1[i] - p2[i]; \
\
sum += t * t; \
} \
\
p1 += sz; \
p2 += lsk; \
} \
\
*q++ = sum; \
} \
} \
}
#define CORR_FLOAT( TYPE ) { \
for( y = 0; y < r->height; y++ ) { \
TYPE *q = (TYPE *) \
VIPS_REGION_ADDR( out, r->left, r->top + y ); \
\
for( x = 0; x < r->width; x++ ) \
for( b = 0; b < bands; b++ ) { \
TYPE *p1 = (TYPE *) ref->data; \
TYPE *p2 = (TYPE *) VIPS_REGION_ADDR( in, \
r->left + x, r->top + y ); \
\
TYPE sum; \
\
sum = 0; \
for( j = 0; j < ref->Ysize; j++ ) { \
for( i = b; i < sz; i += bands ) { \
TYPE t = p1[i] - p2[i]; \
\
sum += t * t; \
} \
\
p1 += sz; \
p2 += lsk; \
} \
\
*q++ = sum; \
} \
} \
}
static void
vips_fastcor_correlation( VipsCorrelation *correlation,
VipsRegion *in, VipsRegion *out )
{
VipsRect *r = &out->valid;
VipsImage *ref = correlation->ref_ready;
int bands = vips_band_format_iscomplex( ref->BandFmt ) ?
ref->Bands * 2 : ref->Bands;
int sz = ref->Xsize * bands;
int lsk = VIPS_REGION_LSKIP( in );
int x, y, i, j, b;
switch( vips_image_get_format( ref ) ) {
case VIPS_FORMAT_CHAR:
CORR_INT( signed char );
break;
case VIPS_FORMAT_UCHAR:
CORR_INT( unsigned char );
break;
case VIPS_FORMAT_SHORT:
CORR_INT( signed short );
break;
case VIPS_FORMAT_USHORT:
CORR_INT( unsigned short );
break;
case VIPS_FORMAT_INT:
CORR_INT( signed int );
break;
case VIPS_FORMAT_UINT:
CORR_INT( unsigned int );
break;
case VIPS_FORMAT_FLOAT:
case VIPS_FORMAT_COMPLEX:
CORR_FLOAT( float );
break;
case VIPS_FORMAT_DOUBLE:
case VIPS_FORMAT_DPCOMPLEX:
CORR_FLOAT( double );
break;
default:
g_assert( 0 );
}
}
/* Save a bit of typing.
*/
#define UC VIPS_FORMAT_UCHAR
#define C VIPS_FORMAT_CHAR
#define US VIPS_FORMAT_USHORT
#define S VIPS_FORMAT_SHORT
#define UI VIPS_FORMAT_UINT
#define I VIPS_FORMAT_INT
#define F VIPS_FORMAT_FLOAT
#define X VIPS_FORMAT_COMPLEX
#define D VIPS_FORMAT_DOUBLE
#define DX VIPS_FORMAT_DPCOMPLEX
/* Type promotion for multiplication. Sign and value preserving. Make sure
* these match the case statement in multiply_buffer() above.
*/
static int vips_fastcor_format_table[10] = {
/* UC C US S UI I F X D DX */
UI, UI, UI, UI, UI, UI,F, X, D, DX
};
static void
vips_fastcor_class_init( VipsFastcorClass *class )
{
VipsObjectClass *object_class = (VipsObjectClass *) class;
VipsCorrelationClass *cclass = VIPS_CORRELATION_CLASS( class );
object_class->nickname = "fastcor";
object_class->description = _( "fast correlation" );
cclass->format_table = vips_fastcor_format_table;
cclass->correlation = vips_fastcor_correlation;
}
static void
vips_fastcor_init( VipsFastcor *fastcor )
{
}
/**
* vips_fastcor:
* @in: input image
* @ref: reference image
* @out: output image
* @...: %NULL-terminated list of optional named arguments
*
* Calculate a fast correlation surface.
*
* @ref is placed at every position in @in and the sum of squares of
* differences calculated.
*
* The output
* image is the same size as the input. Extra input edge pixels are made by
* copying the existing edges outwards.
*
* If the number of bands differs, one of the images
* must have one band. In this case, an n-band image is formed from the
* one-band image by joining n copies of the one-band image together, and then
* the two n-band images are operated upon.
*
* The output type is uint if both inputs are integer, float if both are float
* or complex, and double if either is double or double complex.
* In other words, the output type is just large enough to hold the whole
* range of possible values.
*
* See also: vips_spcor().
*
* Returns: 0 on success, -1 on error
*/
int
vips_fastcor( VipsImage *in, VipsImage *ref, VipsImage **out, ... )
{
va_list ap;
int result;
va_start( ap, out );
result = vips_call_split( "fastcor", ap, in, ref, out );
va_end( ap );
return( result );
}

206
libvips/convolution/im_fastcor.c
View File

@ -1,206 +0,0 @@
/* im_fastcor
*
* Copyright: 1990, N. Dessipris.
*
* Author: Nicos Dessipris
* Written on: 02/05/1990
* Modified on : 15/03/1991
* 20/2/95 JC
* - ANSIfied
* - in1 and in2 swapped, to match order for im_spcor
* - memory leaks fixed
* 21/2/95 JC
* - partialed
* - speed-ups
* 7/4/04
* - now uses im_embed() with edge stretching on the output
* - sets Xoffset / Yoffset
* 8/3/06 JC
* - use im_embed() with edge stretching on the input, not the output
* - calculate sum of squares of differences, rather than abs of
* difference
* 3/2/10
* - gtkdoc
* - cleanups
*/
/*
This file is part of VIPS.
VIPS is free software; you can redistribute it and/or modify
it under the terms of the GNU Lesser General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
02110-1301 USA
*/
/*
These files are distributed with VIPS - http://www.vips.ecs.soton.ac.uk
*/
#ifdef HAVE_CONFIG_H
#include <config.h>
#endif /*HAVE_CONFIG_H*/
#include <vips/intl.h>
#include <stdio.h>
#include <math.h>
#include <vips/vips.h>
/* Fastcor generate function.
*/
static int
fastcor_gen( REGION *or, void *seq, void *a, void *b )
{
REGION *ir = (REGION *) seq;
IMAGE *ref = (IMAGE *) b;
Rect irect;
Rect *r = &or->valid;
int x, y, i, j;
int lsk;
/* What part of ir do we need?
*/
irect.left = or->valid.left;
irect.top = or->valid.top;
irect.width = or->valid.width + ref->Xsize - 1;
irect.height = or->valid.height + ref->Ysize - 1;
if( im_prepare( ir, &irect ) )
return( -1 );
lsk = IM_REGION_LSKIP( ir );
/* Loop over or.
*/
for( y = 0; y < r->height; y++ ) {
unsigned int *q = (unsigned int *)
IM_REGION_ADDR( or, r->left, r->top + y );
for( x = 0; x < r->width; x++ ) {
VipsPel *b = ref->data;
VipsPel *a =
IM_REGION_ADDR( ir, r->left + x, r->top + y );
int sum;
sum = 0;
for( j = 0; j < ref->Ysize; j++ ) {
for( i = 0; i < ref->Xsize; i++ ) {
int t = b[i] - a[i];
sum += t * t;
}
a += lsk;
b += ref->Xsize;
}
q[x] = sum;
}
}
return( 0 );
}
/* Raw fastcor, with no borders.
*/
int
im_fastcor_raw( IMAGE *in, IMAGE *ref, IMAGE *out )
{
/* PIO between in and out; WIO from ref.
*/
if( im_piocheck( in, out ) ||
im_incheck( ref ) )
return( -1 );
/* Check sizes.
*/
if( in->Xsize < ref->Xsize || in->Ysize < ref->Ysize ) {
im_error( "im_fastcor", "%s",
_( "ref not smaller than or equal to in" ) );
return( -1 );
}
/* Check types.
*/
if( im_check_uncoded( "im_fastcor", in ) ||
im_check_mono( "im_fastcor", in ) ||
im_check_format( "im_fastcor", in, IM_BANDFMT_UCHAR ) ||
im_check_coding_same( "im_fastcor", in, ref ) ||
im_check_bands_same( "im_fastcor", in, ref ) ||
im_check_format_same( "im_fastcor", in, ref ) )
return( -1 );
/* Prepare the output image.
*/
if( im_cp_descv( out, in, ref, NULL ) )
return( -1 );
out->BandFmt = IM_BANDFMT_UINT;
out->Xsize = in->Xsize - ref->Xsize + 1;
out->Ysize = in->Ysize - ref->Ysize + 1;
/* FATSTRIP is good for us, as THINSTRIP will cause
* too many recalculations on overlaps.
*/
if( im_demand_hint( out, IM_FATSTRIP, in, NULL ) ||
im_generate( out,
im_start_one, fastcor_gen, im_stop_one, in, ref ) )
return( -1 );
out->Xoffset = -ref->Xsize / 2;
out->Yoffset = -ref->Ysize / 2;
return( 0 );
}
/**
* im_fastcor:
* @in: input image
* @ref: reference image
* @out: output image
*
* Calculate a fast correlation surface.
*
* @ref is placed at every position in @in and the sum of squares of
* differences calculated. One-band, 8-bit unsigned images only. The output
* image is always %IM_BANDFMT_UINT. @ref must be smaller than or equal to
* @in. The output
* image is the same size as the input.
*
* See also: im_spcor().
*
* Returns: 0 on success, -1 on error
*/
int
im_fastcor( IMAGE *in, IMAGE *ref, IMAGE *out )
{
IMAGE *t1 = im_open_local( out, "im_fastcor intermediate", "p" );
if( !t1 ||
im_embed( in, t1, 1,
ref->Xsize / 2, ref->Ysize / 2,
in->Xsize + ref->Xsize - 1,
in->Ysize + ref->Ysize - 1 ) ||
im_fastcor_raw( t1, ref, out ) )
return( -1 );
out->Xoffset = 0;
out->Yoffset = 0;
return( 0 );
}

12
libvips/convolution/im_sharpen.c
View File

@ -96,10 +96,10 @@ typedef struct {
static SharpenLut *
build_lut( IMAGE *out, int x1, int x2, int x3, double m1, double m2 )
{
SharpenLut *slut;
int i;
SharpenLut *slut = IM_NEW( out, SharpenLut );
if( !slut )
if( !(slut = IM_NEW( out, SharpenLut )) )
return( NULL );
if( !(slut->lut = IM_ARRAY( out, x2 + x3 + 1, int )) )
@ -109,13 +109,13 @@ build_lut( IMAGE *out, int x1, int x2, int x3, double m1, double m2 )
slut->x3 = x3;
for( i = 0; i < x1; i++ ) {
slut->lut[x3 + i] = i*m1;
slut->lut[x3 - i] = -i*m1;
slut->lut[x3 + i] = i * m1;
slut->lut[x3 - i] = -i * m1;
}
for( i = x1; i <= x2; i++ )
slut->lut[x3 + i] = x1*m1 + (i-x1)*m2;
slut->lut[x3 + i] = x1 * m1 + (i - x1) * m2;
for( i = x1; i <= x3; i++ )
slut->lut[x3 - i] = -(x1*m1 + (i-x1)*m2);
slut->lut[x3 - i] = -(x1 * m1 + (i - x1) * m2);
return( slut );
}

348
libvips/convolution/im_spcor.c
View File

@ -1,348 +0,0 @@
/* im_spcor
*
* Copyright: 1990, N. Dessipris; 2006, 2007 Nottingham Trent University.
*
*
* Author: Nicos Dessipris
* Written on: 02/05/1990
* Modified on :
* 20/2/95 JC
* - updated
* - ANSIfied, a little
* 21/2/95 JC
* - rewritten
* - partialed
* - speed-ups
* - new correlation coefficient (see above), from Niblack "An
* Introduction to Digital Image Processing", Prentice/Hall, pp 138.
* 4/9/97 JC
* - now does short/ushort as well
* 13/2/03 JC
* - oops, could segv for short images
* 14/4/04 JC
* - sets Xoffset / Yoffset
* 8/3/06 JC
* - use im_embed() with edge stretching on the input, not the output
*
* 2006-10-24 tcv
* - add im_spcor2
*
* 2007-11-12 tcv
* - make im_spcor a wrapper selecting either im__spcor or im__spcor2
* 2008-09-09 JC
* - roll back the windowed version for now, it has some tile edge effects
* 3/2/10
* - gtkdoc
* - cleanups
*/
/*
This file is part of VIPS.
VIPS is free software; you can redistribute it and/or modify
it under the terms of the GNU Lesser General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
02110-1301 USA
*/
/*
These files are distributed with VIPS - http://www.vips.ecs.soton.ac.uk
*/
#ifdef HAVE_CONFIG_H
#include <config.h>
#endif /*HAVE_CONFIG_H*/
#include <vips/intl.h>
#include <stdlib.h>
#include <string.h>
#include <stdio.h>
#include <math.h>
#include <vips/vips.h>
/* Hold per-call state here.
*/
typedef struct {
IMAGE *ref; /* Image we are searching for */
double rmean; /* Mean of search window */
double c1; /* sqrt(sumij (ref(i,j)-mean(ref))^2) */
} Spcor;
#define LOOP( IN ) { \
IN *a = (IN *) p; \
IN *b = (IN *) ref->data; \
int in_lsk = lsk / sizeof( IN ); \
IN *a1; \
IN *b1; \
\
/* For each pel in or, loop over ref. First, \
* calculate mean of area in ir corresponding to ref. \
*/ \
a1 = a; \
sum1 = 0; \
for( j = 0; j < ref->Ysize; j++ ) { \
for( i = 0; i < ref->Xsize; i++ ) \
sum1 += a1[i]; \
a1 += in_lsk; \
} \
imean = sum1 / VIPS_IMAGE_N_PELS( ref ); \
\
/* Loop over ir again, this time calculating \
* sum-of-squares-of-differences for this window on \
* ir, and also sum-of-products-of-differences from mean. \
*/ \
a1 = a; \
b1 = b; \
sum2 = 0.0; \
sum3 = 0.0; \
for( j = 0; j < ref->Ysize; j++ ) { \
for( i = 0; i < ref->Xsize; i++ ) { \
/* Reference pel, and input pel. \
*/ \
IN rp = b1[i]; \
IN ip = a1[i]; \
\
/* Accumulate sum-of-squares-of- \
* differences for input image. \
*/ \
double t = ip - imean; \
sum2 += t * t; \
\
/* Accumulate product-of-difference from mean. \
*/ \
sum3 += (rp - spcor->rmean) * (ip - imean); \
} \
a1 += in_lsk; \
b1 += ref->Xsize; \
} \
}
/* spcor generate function.
*/
static int
spcor_gen( REGION *or, void *vseq, void *a, void *b )
{
REGION *ir = (REGION *) vseq;
Spcor *spcor = (Spcor *) b;
IMAGE *ref = spcor->ref;
Rect irect;
Rect *r = &or->valid;
int le = r->left;
int to = r->top;
int bo = IM_RECT_BOTTOM( r );
int ri = IM_RECT_RIGHT( r );
int x, y, i, j;
int lsk;
double imean;
double sum1;
double sum2, sum3;
double c2, cc;
/* What part of ir do we need?
*/
irect.left = or->valid.left;
irect.top = or->valid.top;
irect.width = or->valid.width + ref->Xsize - 1;
irect.height = or->valid.height + ref->Ysize - 1;
if( im_prepare( ir, &irect ) )
return( -1 );
lsk = IM_REGION_LSKIP( ir );
/* Loop over or.
*/
for( y = to; y < bo; y++ ) {
float *q = (float *) IM_REGION_ADDR( or, le, y );
for( x = le; x < ri; x++ ) {
VipsPel *p = IM_REGION_ADDR( ir, x, y );
/* Find sums for this position.
*/
switch( ref->BandFmt ) {
case IM_BANDFMT_UCHAR: LOOP(unsigned char); break;
case IM_BANDFMT_CHAR: LOOP(signed char); break;
case IM_BANDFMT_USHORT: LOOP(unsigned short); break;
case IM_BANDFMT_SHORT: LOOP(signed short); break;
default:
g_assert( 0 );
/* Keep -Wall happy.
*/
return( 0 );
}
/* Now: calculate correlation coefficient!
*/
c2 = sqrt( sum2 );
cc = sum3 / (spcor->c1 * c2);
*q++ = cc;
}
}
return( 0 );
}
/* Pre-calculate stuff for our reference image.
*/
static Spcor *
spcor_new( IMAGE *out, IMAGE *ref )
{
Spcor *spcor;
guint64 sz = VIPS_IMAGE_N_PELS( ref );
VipsPel *p = ref->data;
double s;
size_t i;
if( !(spcor = IM_NEW( out, Spcor )) )
return( NULL );
/* Pre-calculate stuff on our reference image.
*/
spcor->ref = ref;
if( im_avg( spcor->ref, &spcor->rmean ) )
return( NULL );
/* Find sqrt-of-sum-of-squares-of-differences.
*/
s = 0.0;
for( i = 0; i < sz; i++ ) {
double t = (int) p[i] - spcor->rmean;
s += t * t;
}
spcor->c1 = sqrt( s );
return( spcor );
}
int
im_spcor_raw( IMAGE *in, IMAGE *ref, IMAGE *out )
{
Spcor *spcor;
/* PIO between in and out; WIO from ref, since it's probably tiny.
*/
if( im_piocheck( in, out ) ||
im_incheck( ref ) )
return( -1 );
/* Check sizes.
*/
if( in->Xsize < ref->Xsize ||
in->Ysize < ref->Ysize ) {
im_error( "im_spcor_raw",
"%s", _( "ref not smaller than or equal to in" ) );
return( -1 );
}
/* Check types.
*/
if( im_check_uncoded( "im_spcor", in ) ||
im_check_mono( "im_spcor", in ) ||
im_check_8or16( "im_spcor", in ) ||
im_check_coding_same( "im_spcor", in, ref ) ||
im_check_bands_same( "im_spcor", in, ref ) ||
im_check_format_same( "im_spcor", in, ref ) )
return( -1 );
/* Prepare the output image.
*/
if( im_cp_descv( out, in, ref, NULL ) )
return( -1 );
out->BandFmt = IM_BANDFMT_FLOAT;
out->Xsize = in->Xsize - ref->Xsize + 1;
out->Ysize = in->Ysize - ref->Ysize + 1;
/* Pre-calculate some stuff.
*/
if( !(spcor = spcor_new( out, ref )) )
return( -1 );
/* Set demand hints. FATSTRIP is good for us, as THINSTRIP will cause
* too many recalculations on overlaps.
*/
if( im_demand_hint( out, IM_FATSTRIP, in, NULL ) )
return( -1 );
/* Write the correlation.
*/
if( im_generate( out,
im_start_one, spcor_gen, im_stop_one, in, spcor ) )
return( -1 );
out->Xoffset = -ref->Xsize / 2;
out->Yoffset = -ref->Ysize / 2;
return( 0 );
}
/**
* im_spcor:
* @in: input image
* @ref: reference image
* @out: output image
*
* Calculate a correlation surface.
*
* @ref is placed at every position in @in and the correlation coefficient
* calculated. One-band, 8 or 16-bit images only. @in and @ref must have the
* same #VipsBandFmt. The output
* image is always %IM_BANDFMT_FLOAT. @ref must be smaller than or equal to
* @in. The output
* image is the same size as the input.
*
* The correlation coefficient is calculated as:
*
* |[
* sumij (ref(i,j)-mean(ref))(inkl(i,j)-mean(inkl))
* c(k,l) = ------------------------------------------------
* sqrt(sumij (ref(i,j)-mean(ref))^2) *
* sqrt(sumij (inkl(i,j)-mean(inkl))^2)
* ]|
*
* where inkl is the area of @in centred at position (k,l).
*
* from Niblack "An Introduction to Digital Image Processing",
* Prentice/Hall, pp 138.
*
* See also: im_gradcor(), im_fastcor().
*
* Returns: 0 on success, -1 on error
*/
int
im_spcor( IMAGE *in, IMAGE *ref, IMAGE *out )
{
IMAGE *t1 = im_open_local( out, "im_spcor intermediate", "p" );
if( !t1 ||
im_embed( in, t1, 1,
ref->Xsize / 2, ref->Ysize / 2,
in->Xsize + ref->Xsize - 1,
in->Ysize + ref->Ysize - 1 ) ||
im_spcor_raw( t1, ref, out ) )
return( -1 );
out->Xoffset = 0;
out->Yoffset = 0;
return( 0 );
}

367
libvips/convolution/spcor.c Normal file
View File

@ -0,0 +1,367 @@
/* spcor
*
* Copyright: 1990, N. Dessipris; 2006, 2007 Nottingham Trent University.
*
*
* Author: Nicos Dessipris
* Written on: 02/05/1990
* Modified on :
* 20/2/95 JC
* - updated
* - ANSIfied, a little
* 21/2/95 JC
* - rewritten
* - partialed
* - speed-ups
* - new correlation coefficient (see above), from Niblack "An
* Introduction to Digital Image Processing", Prentice/Hall, pp 138.
* 4/9/97 JC
* - now does short/ushort as well
* 13/2/03 JC
* - oops, could segv for short images
* 14/4/04 JC
* - sets Xoffset / Yoffset
* 8/3/06 JC
* - use im_embed() with edge stretching on the input, not the output
*
* 2006-10-24 tcv
* - add im_spcor2
*
* 2007-11-12 tcv
* - make im_spcor a wrapper selecting either im__spcor or im__spcor2
* 2008-09-09 JC
* - roll back the windowed version for now, it has some tile edge effects
* 3/2/10
* - gtkdoc
* - cleanups
* 7/11/13
* - redone as a class
*/
/*
This file is part of VIPS.
VIPS is free software; you can redistribute it and/or modify
it under the terms of the GNU Lesser General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
02110-1301 USA
*/
/*
These files are distributed with VIPS - http://www.vips.ecs.soton.ac.uk
*/
#ifdef HAVE_CONFIG_H
#include <config.h>
#endif /*HAVE_CONFIG_H*/
#include <vips/intl.h>
#include <stdio.h>
#include <math.h>
#include <vips/vips.h>
#include "pconvolution.h"
#include "correlation.h"
typedef struct _VipsSpcor {
VipsCorrelation parent_instance;
/* Per-band mean of ref images.
*/
double *rmean;
/* Per band sqrt(sumij (ref(i,j)-mean(ref))^2)
*/
double *c1;
} VipsSpcor;
typedef VipsCorrelationClass VipsSpcorClass;
G_DEFINE_TYPE( VipsSpcor, vips_spcor, VIPS_TYPE_CORRELATION );
static int
vips_spcor_pre_generate( VipsCorrelation *correlation )
{
VipsObjectClass *class = VIPS_OBJECT_GET_CLASS( correlation );
VipsSpcor *spcor = (VipsSpcor *) correlation;
VipsImage *ref = correlation->ref_ready;
int bands = ref->Bands;
VipsImage **b = (VipsImage **)
vips_object_local_array( VIPS_OBJECT( spcor ), bands );
VipsImage **t = (VipsImage **)
vips_object_local_array( VIPS_OBJECT( spcor ), 2 );
VipsImage **b2 = (VipsImage **)
vips_object_local_array( VIPS_OBJECT( spcor ), bands );
int i;
double *offset;
double *scale;
if( vips_check_noncomplex( class->nickname, ref ) )
return( -1 );
/* Per-band mean.
*/
if( !(spcor->rmean = VIPS_ARRAY( spcor, bands, double )) ||
!(spcor->c1 = VIPS_ARRAY( spcor, bands, double )) )
return( -1 );
for( i = 0; i < bands; i++ )
if( vips_extract_band( ref, &b[i], i, NULL ) ||
vips_avg( b[i], &spcor->rmean[i], NULL ) )
return( -1 );
/* Per band sqrt(sumij (ref(i,j)-mean(ref))^2)
*/
if( !(offset = VIPS_ARRAY( spcor, bands, double )) ||
!(scale = VIPS_ARRAY( spcor, bands, double )) )
return( -1 );
for( i = 0; i < bands; i++ ) {
offset[i] = -spcor->rmean[i];
scale[i] = 1.0;
}
if( vips_linear( ref, &t[0], scale, offset, bands, NULL ) ||
vips_multiply( t[0], t[0], &t[1], NULL ) )
return( -1 );
for( i = 0; i < bands; i++ )
if( vips_extract_band( t[1], &b2[i], i, NULL ) ||
vips_avg( b2[i], &spcor->c1[i], NULL ) )
return( -1 );
for( i = 0; i < bands; i++ ) {
spcor->c1[i] *= ref->Xsize * ref->Ysize;
spcor->c1[i] = sqrt( spcor->c1[i] );
}
return( 0 );
}
#define LOOP( IN ) { \
IN *r1 = ((IN *) ref->data) + b; \
IN *p1 = ((IN *) p) + b; \
int in_lsk = lsk / sizeof( IN ); \
IN *r1a; \
IN *p1a; \
\
/* Mean of area of in corresponding to ref. \
*/ \
p1a = p1; \
sum1 = 0.0; \
for( j = 0; j < ref->Ysize; j++ ) { \
for( i = 0; i < sz; i += bands ) \
sum1 += p1a[i]; \
p1a += in_lsk; \
} \
imean = sum1 / VIPS_IMAGE_N_PELS( ref ); \
\
/* Calculate sum-of-squares-of-differences for this window on \
* in, and also sum-of-products-of-differences from mean. \
*/ \
p1a = p1; \
r1a = r1; \
sum2 = 0.0; \
sum3 = 0.0; \
for( j = 0; j < ref->Ysize; j++ ) { \
for( i = 0; i < sz; i += bands ) { \
/* Reference pel, and input pel. \
*/ \
IN ip = p1a[i]; \
IN rp = r1a[i]; \
\
/* Accumulate sum-of-squares-of- \
* differences for input image. \
*/ \
double t = ip - imean; \
sum2 += t * t; \
\
/* Accumulate product-of-difference from mean. \
*/ \
sum3 += (rp - spcor->rmean[b]) * (ip - imean); \
} \
\
p1a += in_lsk; \
r1a += sz; \
} \
}
static void
vips_spcor_correlation( VipsCorrelation *correlation,
VipsRegion *in, VipsRegion *out )
{
VipsSpcor *spcor = (VipsSpcor *) correlation;
VipsRect *r = &out->valid;
VipsImage *ref = correlation->ref_ready;
int bands = vips_band_format_iscomplex( ref->BandFmt ) ?
ref->Bands * 2 : ref->Bands;
int sz = ref->Xsize * bands;
int lsk = VIPS_REGION_LSKIP( in );
int x, y, b, j, i;
double imean;
double sum1;
double sum2, sum3;
double c2, cc;
for( y = 0; y < r->height; y++ ) {
float *q = (float *)
VIPS_REGION_ADDR( out, r->left, r->top + y );
for( x = 0; x < r->width; x++ ) {
VipsPel *p =
VIPS_REGION_ADDR( in, r->left + x, r->top + y );
for( b = 0; b < bands; b++ ) {
switch( vips_image_get_format( ref ) ) {
case VIPS_FORMAT_UCHAR:
LOOP( unsigned char );
break;
case VIPS_FORMAT_CHAR:
LOOP( signed char );
break;
case VIPS_FORMAT_USHORT:
LOOP( unsigned short );
break;
case VIPS_FORMAT_SHORT:
LOOP( signed short );
break;
case VIPS_FORMAT_UINT:
LOOP( unsigned int );
break;
case VIPS_FORMAT_INT:
LOOP( signed int );
break;
case VIPS_FORMAT_FLOAT:
case VIPS_FORMAT_COMPLEX:
LOOP( float );
break;
case VIPS_FORMAT_DOUBLE:
case VIPS_FORMAT_DPCOMPLEX:
LOOP( double );
break;
default:
g_assert( 0 );
}
c2 = sqrt( sum2 );
cc = sum3 / (spcor->c1[b] * c2);
*q++ = cc;
}
}
}
}
/* Save a bit of typing.
*/
#define UC VIPS_FORMAT_UCHAR
#define C VIPS_FORMAT_CHAR
#define US VIPS_FORMAT_USHORT
#define S VIPS_FORMAT_SHORT
#define UI VIPS_FORMAT_UINT
#define I VIPS_FORMAT_INT
#define F VIPS_FORMAT_FLOAT
#define X VIPS_FORMAT_COMPLEX
#define D VIPS_FORMAT_DOUBLE
#define DX VIPS_FORMAT_DPCOMPLEX
static int vips_spcor_format_table[10] = {
/* UC C US S UI I F X D DX */
F, F, F, F, F, F, F, F, F, F
};
static void
vips_spcor_class_init( VipsSpcorClass *class )
{
VipsObjectClass *object_class = (VipsObjectClass *) class;
VipsCorrelationClass *cclass = VIPS_CORRELATION_CLASS( class );
object_class->nickname = "spcor";
object_class->description = _( "spatial correlation" );
cclass->format_table = vips_spcor_format_table;
cclass->pre_generate = vips_spcor_pre_generate;
cclass->correlation = vips_spcor_correlation;
}
static void
vips_spcor_init( VipsSpcor *spcor )
{
}
/**
* vips_spcor:
* @in: input image
* @ref: reference image
* @out: output image
* @...: %NULL-terminated list of optional named arguments
*
* Calculate a correlation surface.
*
* @ref is placed at every position in @in and the correlation coefficient
* calculated. The output
* image is always float.
*
* The output
* image is the same size as the input. Extra input edge pixels are made by
* copying the existing edges outwards.
*
* The correlation coefficient is calculated as:
*
* |[
* sumij (ref(i,j)-mean(ref))(inkl(i,j)-mean(inkl))
* c(k,l) = ------------------------------------------------
* sqrt(sumij (ref(i,j)-mean(ref))^2) *
* sqrt(sumij (inkl(i,j)-mean(inkl))^2)
* ]|
*
* where inkl is the area of @in centred at position (k,l).
*
* from Niblack "An Introduction to Digital Image Processing",
* Prentice/Hall, pp 138.
*
* If the number of bands differs, one of the images
* must have one band. In this case, an n-band image is formed from the
* one-band image by joining n copies of the one-band image together, and then
* the two n-band images are operated upon.
*
* The output image is always float, unless either of the two inputs is
* double, in which case the output is also double.
*
* See also: vips_gradcor(), vips_fastcor().
*
* Returns: 0 on success, -1 on error
*/
int
vips_spcor( VipsImage *in, VipsImage *ref, VipsImage **out, ... )
{
va_list ap;
int result;
va_start( ap, out );
result = vips_call_split( "spcor", ap, in, ref, out );
va_end( ap );
return( result );
}

2
libvips/deprecated/Makefile.am
View File

@ -9,6 +9,7 @@ libdeprecated_la_SOURCES = \
im_lab_morph.c \
deprecated_dispatch.c \
colour_dispatch.c \
convol_dispatch.c \
arith_dispatch.c \
hist_dispatch.c \
im_maxpos_avg.c \
@ -24,6 +25,7 @@ libdeprecated_la_SOURCES = \
im_fav4.c \
im_gadd.c \
im_gaddim.c \
im_gradcor.c \
im_cmulnorm.c \
im_printlines.c \
im_convsub.c \

0
libvips/convolution/convol_dispatch.c → libvips/deprecated/convol_dispatch.c
View File

0
libvips/convolution/im_gradcor.c → libvips/deprecated/im_gradcor.c
View File

46
libvips/deprecated/vips7compat.c
View File

@ -2278,6 +2278,52 @@ im_contrast_surface( IMAGE *in, IMAGE *out, int half_win_size, int spacing )
return( 0 );
}
int
im_spcor_raw( IMAGE *in, IMAGE *ref, IMAGE *out )
{
im_error( "im_spcor_raw", "no compat function" );
return( -1 );
}
int
im_spcor( IMAGE *in, IMAGE *ref, IMAGE *out )
{
VipsImage *x;
if( vips_call( "spcor", in, ref, &x, NULL ) )
return( -1 );
if( im_copy( x, out ) ) {
g_object_unref( x );
return( -1 );
}
g_object_unref( x );
return( 0 );
}
int
im_fastcor_raw( IMAGE *in, IMAGE *ref, IMAGE *out )
{
im_error( "im_fastcor_raw", "no compat function" );
return( -1 );
}
int
im_fastcor( IMAGE *in, IMAGE *ref, IMAGE *out )
{
VipsImage *x;
if( vips_call( "fastcor", in, ref, &x, NULL ) )
return( -1 );
if( im_copy( x, out ) ) {
g_object_unref( x );
return( -1 );
}
g_object_unref( x );
return( 0 );
}
static int
vips__round( VipsImage *in, VipsImage *out, VipsOperationRound round )
{

1
libvips/foreign/radiance.c
View File

@ -788,7 +788,6 @@ scanline_write( COLR *scanline, int width, FILE *fp )
}
int i, j, beg, cnt;
int c2;
if( width < MINELEN ||
width > MAXELEN )

7
libvips/include/vips/convolution.h
View File

@ -88,13 +88,6 @@ int im_sharpen( VipsImage *in, VipsImage *out,
double x1, double y2, double y3,
double m1, double m2 );
int im_grad_x( VipsImage *in, VipsImage *out );
int im_grad_y( VipsImage *in, VipsImage *out );
int im_fastcor( VipsImage *in, VipsImage *ref, VipsImage *out );
int im_spcor( VipsImage *in, VipsImage *ref, VipsImage *out );
int im_gradcor( VipsImage *in, VipsImage *ref, VipsImage *out );
#ifdef __cplusplus
}
#endif /*__cplusplus*/

7
libvips/include/vips/vips7compat.h
View File

@ -923,6 +923,13 @@ int im_contrast_surface_raw( IMAGE *in, IMAGE *out,
int im_contrast_surface( VipsImage *in, VipsImage *out,
int half_win_size, int spacing );
int im_grad_x( VipsImage *in, VipsImage *out );
int im_grad_y( VipsImage *in, VipsImage *out );
int im_fastcor( VipsImage *in, VipsImage *ref, VipsImage *out );
int im_spcor( VipsImage *in, VipsImage *ref, VipsImage *out );
int im_gradcor( VipsImage *in, VipsImage *ref, VipsImage *out );
#ifdef __cplusplus
}
#endif /*__cplusplus*/