1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347
|
/* warp.c : Digital Image Warping algorithms
//
// See George Wolberg's "Digital Image Warping"
// IEEE Computer Society Press order number 1944
// ISBN 0-8186-8944-7
//
// Copyrights might be held by various authors. See individual routines.
//
Written and Copyright (C) 1994-1999 by Michael J. Gourlay
This file is part of Xmorph.
Xmorph is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2, or (at your option)
any later version.
Xmorph 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 General Public License for more details.
You should have received a copy of the GNU General Public License
along with Xmorph; see the file LICENSE. If not, write to
the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
Boston, MA 02111-1307, USA.
*/
#include <stdio.h>
#include <stdlib.h>
#include "my_malloc.h"
#include "spline.h"
#include "spl-array.h"
#include "warp.h"
#define MAX(x,y) ((x)>(y) ? (x) : (y))
#include "resample.h"
/* --------------------------------------------------------------- */
/* resample_array: Fant's resampling algorithm, specialized to
// images with the same dimensions
//
// Based on "Nonaliasing Real-Time Spatial Transform" by Karl M. Fant.
// See IEEE Computer Graphics and Applications v6n1, January 1986.
//
// NOTE: The original code contained array bounds overruns. I fixed
// the code up without taking resampling theory into consideration. I
// figured the boundary pixels would be the only things hosed anyway, and
// the original code was dumping core. -- MJG 08jun94
//
// UPDATE: 20jul94: George Wolberg (himself) gives these array bounds
// modifications a thumbs up.
*/
static void
resample_array(const double *F, const PIXEL_TYPE *src, PIXEL_TYPE *dst, int len, int ipo)
{
int ui, xi;
double acc, intensity;
double INSFAC, INSEG, OUTSEG;
double *inpos; /* forward mapping function */
/* MJG 08jun94 */
inpos = MY_CALLOC(len+2, double);
/* precompute input index for each output pixel */
for(ui=xi=0; xi<len; xi++) {
while( ui<(len-1) && F[ui+1]<xi ) ui++; /* MJG 08jun94 */
if(ui<(len-1)) { /* MJG 08jun94 */
inpos[xi] = ui + (double) (xi - F[ui]) / (F[ui+1] - F[ui]);
} else {
inpos[xi] = ui + 1.0 ;
}
}
inpos[len] = len; /* MJG 08jun94 */
INSEG = 1.0; /* entire input pixel is available */
OUTSEG = inpos[1]; /* # input pixels that map onto 1 output pixel */
INSFAC = OUTSEG; /* inverse scale factor */
acc = 0.0; /* accumulator */
/* compute all output pixels */
/* ellenberger@tle.enet.dec.com reports:
// the loop upper bound should be (len - 1) instead of len
// Just in case, I'm changing it to (len - 1)
// Consider this a hack until I figure out what's going on. -- 13sep95 MJG
*/
/* for(xi = ui = 0; xi < len; ) { */
for(xi = ui = 0; xi < (len - 1); ) {
#ifdef ARRAY_CHECK
if(ui > ((len-1)*ipo)) {
fprintf(stderr, "resample_array: index out of range: %i\n", ui);
}
#endif
/* use linear interpolation for reconstruction */
intensity = (INSEG * src[ui]) + ((1-INSEG) * src[ui+1]);
/* INSEG < OUTSEG: input pixel is entirely consumed before output pixel */
if(INSEG < OUTSEG) {
acc += (intensity * INSEG); /* accumulation of weighted contrib */
OUTSEG -= INSEG; /* INSEG portion has been filled */
INSEG = 1.0; /* new input pixel will be available */
ui += ipo; /* index to next input pixel */
}
/* INSEG >= OUTSEG: input pixel is not consumed before output pixel */
else {
acc += (intensity * OUTSEG); /* accumulate weighted contrib */
/* A short hack to prevent an arithmetic exception on division by zero
// in case INSFAC becomes zero (AT 04feb95) -- Added 13sep95
// [It seems that when INSFAC is zero, so is acc -- MJG 13sep95]
*/
if (INSFAC==0) {
INSFAC = 1;
}
/* End of hack (AT 04feb95) */
dst[xi*ipo] = acc/INSFAC; /* init output w/ normalized accumulator */
acc = 0.0; /* reset accumulator for next output pixel */
INSEG -= OUTSEG; /* OUTSEG portion of input has been used */
xi++; /* index to next output pixel */
INSFAC = inpos[xi+1] - inpos[xi]; /* init spatially varying INSFAC */
OUTSEG = INSFAC; /* init spatially varying SIZFAC */
}
}
FREE(inpos);
}
/* --------------------------------------------------------------- */
/* NAME
// warp_image: 2-pass mesh-based single-channel image warping
//
//
// ARGUMENTS
// in (in) : input image (single channel, 1 byte per pixel)
// out (out) : output image (single channel, 1 byte per pixel)
// img_width (in) : width dimension, in pixels, of in and out images
// img_height (in) : height dimension, in pixels, of in and out images
// xs (in) : source mesh x-coordinate values
// ys (in) : source mesh y-coordinate values
// xd (in) : destination mesh x-coordinate values
// yd (in) : destination mesh y-coordinate values
// mesh_width (in) : x-dimension of meshes
// mesh_height (in): y-dimension of meshes
//
//
// NOTES
// Mesh points must not fold over,
// and border mesh points must stay on border
//
// Either this routine or the resample routine is responsible for an
// error where the right vertical line and the bottom horizontal line are
// entirely black. This is probably due to a descrepency somewhere
// between the "nx and ny" of an image, and the last image index in each
// direction, which is (nx-1, ny-1). Some day I will track down this bog
// and fix it. Meanwhile, remember to crop the final image by 1 pixel in
// each direction. (6jun97 MJG)
//
// A Mennucc: I have had a striking idea!!!
// I have inverted the roles of src & dst!!!
// so now the resample_array becomes resample_array_inv!!!
// which now pulls pixels into dst instead of pushing from src.
// So the above notes are probably false now. (nov 03)
//
// AUTHOR
// This code was originally written by George Wolberg, based on Smythe90.
// Modifications and many bug fixes by Michael J. Gourlay and other
// authors as noted in the code comments.
//
//
// SEE ALSO
// See Douglas B. Smythe "A Two-Pass Mesh Warping Algorithm for
// Object Transformation and Image Interpolation", ILM Technical Memo
// #1030, Computer Graphics Department, Lucasfilm Ltd., 1990
*/
void
warp_image_inv_old(const PIXEL_TYPE *in, PIXEL_TYPE *out, int img_width, int img_height,
const double *xs, const double *ys, const double *xd, const double *yd,
int mesh_width, int mesh_height)
{
/* "const type *" means "pointer to a constant array of type"
// not "constant pointer to type"
*/
int ai;
int xi, yi;
const PIXEL_TYPE *src;
PIXEL_TYPE *dst;
PIXEL_TYPE *tmp;
const double *x1, *y1, *x2, *y2;
double *xrow1, *yrow1, *xrow2, *yrow2;
double *map1, *map2;
double *indx;
double *ts, *ti, *td;
/* allocate memory for buffers:
//
// indx stores indices used to sample splines
// xrow1, xrow2, yrow2, yrow2 store column data in row order for spline
// map1, map2 store mapping functions computed in row order in spline
//
// Could use alloca instead to avoid the free's at the end of this
// routine. alloca might also be faster. In fact, the original code
// might have used alloca. I do not remember why I changed it.
*/
ai = MAX(img_height, img_width) + 1;
indx = MY_CALLOC(ai, double);
xrow1 = MY_CALLOC(ai, double);
yrow1 = MY_CALLOC(ai, double);
xrow2 = MY_CALLOC(ai, double);
yrow2 = MY_CALLOC(ai, double);
map1 = MY_CALLOC(ai, double);
map2 = MY_CALLOC(ai, double);
/* First pass (phase one): create tables ts and ti for x-intercepts
// of vertical splines in S and I.
// tables have mesh_width columns of height img_height
*/
ts = MY_CALLOC(mesh_width*img_height, double);
ti = MY_CALLOC(mesh_width*img_height, double);
for(yi=0; yi<img_height; yi++) /* indices used to sample vert spline */
indx[yi] = yi;
for(xi=0; xi<mesh_width; xi++) { /* visit each vertical spline */
/* store columns as rows for spline */
for(yi=0; yi<mesh_height; yi++) {
xrow1[yi] = xs[yi*mesh_width+xi];
xrow2[yi] = xd[yi*mesh_width+xi];
yrow1[yi] = ys[yi*mesh_width+xi];
yrow2[yi] = yd[yi*mesh_width+xi];
}
/* scan convert vertical splines of S and I */
hermite3_array(yrow1, xrow1, mesh_height, indx, map1, img_height);
/* bug reported 30may96 by Mike Hoefelein: */
/* hermite3_array(yrow2, xrow2, mesh_height, indx, map2, img_height); */
hermite3_array(yrow1, xrow2, mesh_height, indx, map2, img_height);
/* store resampled rows back into columns */
for(yi=0; yi<img_height; yi++) {
ts[yi*mesh_width+xi] = map1[yi];
ti[yi*mesh_width+xi] = map2[yi];
}
}
/* First pass (phase two): warp x using ts and ti.
// tmp holds intermediate image.
*/
if((tmp = MY_CALLOC(img_height * img_width, PIXEL_TYPE))==NULL) {
fprintf(stderr, "warp_image: Bad Alloc: tmp\n"); return;
}
for(xi=0; xi < img_width; xi++) /* indices used to sample horiz spline */
indx[xi]=xi;
for(yi=0; yi < img_height; yi++) { /* visit each row */
/* fit spline to x-intercepts; resample over all columns */
x1 = &ts[yi*mesh_width];
x2 = &ti[yi*mesh_width];
hermite3_array(x2, x1, mesh_width, indx, map1, img_width);
/* resample source row based on map1 */
src = &in[yi*img_width];
dst = &tmp[yi*img_width];
resample_array_inv_bc(map1, src, dst, img_width, 1);
}
/* free buffers */
FREE(ts);
FREE(ti);
/* Second pass (phase one): create tables ti and td for y-intercepts
// of horiz splines in I and D.
// Tables have mesh_height rows of width img_width
*/
ti = MY_CALLOC(mesh_height*img_width, double);
td = MY_CALLOC(mesh_height*img_width, double);
for(xi=0; xi < img_width; xi++)
indx[xi] = xi; /* indices used to sample horiz splines */
for(yi=0; yi < mesh_height; yi++) {
/* scan convert horizontal splines of I and D */
#if 0
x1 = &xs[yi*mesh_width];
#endif
y1 = &ys[yi*mesh_width];
x2 = &xd[yi*mesh_width];
y2 = &yd[yi*mesh_width];
/* The following line is correct: (x2, y1, ...) */
hermite3_array(x2, y1, mesh_width, indx, &ti[yi*img_width], img_width);
hermite3_array(x2, y2, mesh_width, indx, &td[yi*img_width], img_width);
}
/* Second pass (phase two): warp y using ti and td */
for(yi=0; yi < img_height; yi++)
indx[yi] = yi;
for(xi=0; xi < img_width; xi++) {
/* store columns as row for hermite3_array */
for(yi=0; yi<mesh_height; yi++) {
xrow1[yi] = ti[yi*img_width+xi];
yrow1[yi] = td[yi*img_width+xi];
}
/* fit spline to y-intercepts: resample over all rows */
hermite3_array(yrow1, xrow1, mesh_height, indx, map1, img_height);
/* resample intermediate image column based on map */
src = &tmp[xi];
dst = &out[xi];
resample_array_inv_bc(map1, src, dst, img_height, img_width);
}
FREE(tmp);
FREE(ti);
FREE(td);
FREE(indx);
FREE(xrow1);
FREE(yrow1);
FREE(xrow2);
FREE(yrow2);
FREE(map1);
FREE(map2);
}
|