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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931
|
/* ----------------------------- MNI Header -----------------------------------
@NAME : mincmath
@INPUT : argc, argv - command line arguments
@OUTPUT : (none)
@RETURNS : status
@DESCRIPTION: Program to do simple math on minc files
@METHOD :
@GLOBALS :
@CALLS :
@CREATED : April 28, 1995 (Peter Neelin)
@MODIFIED :
* $Log: mincmath.c,v $
* Revision 6.15 2009-01-20 11:58:13 rotor
* * CMakeLists.txt: updated version
* * Updated Changelog to include releases
* * Warning cleanups below
* * conversion/dcm2mnc/minc_file.c: fixed printf type
* * conversion/dcm2mnc/siemens_to_dicom.c: fixed printf type
* * conversion/ecattominc/machine_indep.c: added string.h and fixed
* 2 fprintf missing format args
* * conversion/micropet/upet2mnc.c: fixed two fprintf format args
* * conversion/minctoecat/ecat_write.c: added string.h
* * conversion/minctoecat/minctoecat.c: added missing argument to fprintf
* * conversion/nifti1/mnc2nii.c: fixed incorrect printf type
* * progs/mincview/invert_raw_image.c: added fwrite checking
*
* Revision 6.14 2008/01/17 02:33:02 rotor
* * removed all rcsids
* * removed a bunch of ^L's that somehow crept in
* * removed old (and outdated) BUGS file
*
* Revision 6.13 2008/01/13 04:30:28 stever
* Add braces around static initializers.
*
* Revision 6.12 2008/01/12 19:08:15 stever
* Add __attribute__ ((unused)) to all rcsid variables.
*
* Revision 6.11 2007/12/11 12:43:01 rotor
* * added static to all global variables in main programs to avoid linking
* problems with libraries (compress in mincconvert and libz for example)
*
* Revision 6.10 2005/08/26 21:07:17 bert
* Use #if rather than #ifdef with MINC2 symbol, and be sure to include config.h whereever MINC2 is used
*
* Revision 6.9 2004/12/14 23:40:07 bert
* Get rid of c99 compilation problems
*
* Revision 6.8 2004/12/03 21:56:51 bert
* Include config.h
*
* Revision 6.7 2004/11/01 22:38:39 bert
* Eliminate all references to minc_def.h
*
* Revision 6.6 2004/04/27 15:31:45 bert
* Added -2 option
*
* Revision 6.5 2001/04/24 13:38:44 neelin
* Replaced NC_NAT with MI_ORIGINAL_TYPE.
*
* Revision 6.4 2001/04/17 18:40:22 neelin
* Modifications to work with NetCDF 3.x
* In particular, changed NC_LONG to NC_INT (and corresponding longs to ints).
* Changed NC_UNSPECIFIED to NC_NAT.
* A few fixes to the configure script.
*
* Revision 6.3 2001/01/15 14:58:43 neelin
* Modified description of -segment option.
*
* Revision 6.2 2000/07/07 13:19:54 neelin
* Added option -filelist to read file names from a file. This gets around
* command-line length limits.
*
* Revision 6.1 1999/10/19 14:45:26 neelin
* Fixed Log subsitutions for CVS
*
* Revision 6.0 1997/09/12 13:24:17 neelin
* Release of minc version 0.6
*
* Revision 5.0 1997/08/21 13:25:16 neelin
* Release of minc version 0.5
*
* Revision 4.0 1997/05/07 20:02:04 neelin
* Release of minc version 0.4
*
* Revision 3.5 1997/04/24 13:48:51 neelin
* Fixed handling of invalid or uninitialized data for cumulative operations.
* Added options -illegal_value and -count_valid.
*
* Revision 3.4 1997/04/23 19:34:56 neelin
* Added options -maximum, -minimum, -abs.
*
* Revision 3.3 1996/01/17 21:24:06 neelin
* Added -exp and -log options.
*
* Revision 3.2 1996/01/16 13:29:31 neelin
* Added -invert option.
*
* Revision 3.1 1995/12/13 16:22:24 neelin
* Added -check_dimensions and -nocheck_dimensions options.
*
* Revision 3.0 1995/05/15 19:32:42 neelin
* Release of minc version 0.3
*
* Revision 1.2 1995/05/03 16:13:46 neelin
* Changed default for -copy/-nocopy to depend on number of input files.
*
* Revision 1.1 1995/05/03 13:19:56 neelin
* Initial revision
*
---------------------------------------------------------------------------- */
#define _GNU_SOURCE 1
#if HAVE_CONFIG_H
#include "config.h"
#endif
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <float.h>
#include <limits.h>
#include <math.h>
#include <minc.h>
#include <ParseArgv.h>
#include <time_stamp.h>
#include <voxel_loop.h>
#include "read_file_names.h"
/* Constants */
#ifndef TRUE
# define TRUE 1
# define FALSE 0
#endif
/* Data values for invalid data and for uninitialized data */
#define INVALID_DATA -DBL_MAX
#define UNINITIALIZED_DATA DBL_MAX
/* Values for representing default case for command-line options */
#define DEFAULT_DBL DBL_MAX
#define DEFAULT_BOOL -1
/* Typedefs */
typedef enum {
UNSPECIFIED_OP = 0, ADD_OP, SUB_OP, MULT_OP, DIV_OP, SQRT_OP, SQUARE_OP,
SCALE_OP, CLAMP_OP, SEGMENT_OP, NSEGMENT_OP, PERCENTDIFF_OP,
EQ_OP, NE_OP, GT_OP, GE_OP, LT_OP, LE_OP, AND_OP, OR_OP, NOT_OP,
ISNAN_OP, NISNAN_OP, INVERT_OP, EXP_OP, LOG_OP, MAX_OP, MIN_OP, ABS_OP,
COUNT_OP
} Operation;
typedef enum {
ILLEGAL_NUMOP, UNARY_NUMOP, BINARY_NUMOP, NARY_NUMOP
} Num_Operands;
/* Table that matches [Operation][Number of constants (0,1,2)] to a number
of volume operands */
Num_Operands OperandTable[][3] = {
{ ILLEGAL_NUMOP, ILLEGAL_NUMOP, ILLEGAL_NUMOP }, /* UNSPECIFIED_OP */
{ NARY_NUMOP, UNARY_NUMOP, ILLEGAL_NUMOP }, /* ADD_OP */
{ BINARY_NUMOP, UNARY_NUMOP, ILLEGAL_NUMOP }, /* SUB_OP */
{ NARY_NUMOP, UNARY_NUMOP, ILLEGAL_NUMOP }, /* MULT_OP */
{ BINARY_NUMOP, UNARY_NUMOP, ILLEGAL_NUMOP }, /* DIV_OP */
{ UNARY_NUMOP, ILLEGAL_NUMOP, ILLEGAL_NUMOP }, /* SQRT_OP */
{ UNARY_NUMOP, ILLEGAL_NUMOP, ILLEGAL_NUMOP }, /* SQUARE_OP */
{ ILLEGAL_NUMOP, UNARY_NUMOP, UNARY_NUMOP }, /* SCALE_OP */
{ ILLEGAL_NUMOP, ILLEGAL_NUMOP, UNARY_NUMOP }, /* CLAMP_OP */
{ ILLEGAL_NUMOP, ILLEGAL_NUMOP, UNARY_NUMOP }, /* SEGMENT_OP */
{ ILLEGAL_NUMOP, ILLEGAL_NUMOP, UNARY_NUMOP }, /* NSEGMENT_OP */
{ BINARY_NUMOP, BINARY_NUMOP, ILLEGAL_NUMOP }, /* PERCENTDIFF_OP */
{ BINARY_NUMOP, UNARY_NUMOP, ILLEGAL_NUMOP }, /* EQ_OP */
{ BINARY_NUMOP, UNARY_NUMOP, ILLEGAL_NUMOP }, /* NE_OP */
{ BINARY_NUMOP, UNARY_NUMOP, ILLEGAL_NUMOP }, /* GT_OP */
{ BINARY_NUMOP, UNARY_NUMOP, ILLEGAL_NUMOP }, /* GE_OP */
{ BINARY_NUMOP, UNARY_NUMOP, ILLEGAL_NUMOP }, /* LT_OP */
{ BINARY_NUMOP, UNARY_NUMOP, ILLEGAL_NUMOP }, /* LE_OP */
{ NARY_NUMOP, ILLEGAL_NUMOP, ILLEGAL_NUMOP }, /* AND_OP */
{ NARY_NUMOP, ILLEGAL_NUMOP, ILLEGAL_NUMOP }, /* OR_OP */
{ UNARY_NUMOP, ILLEGAL_NUMOP, ILLEGAL_NUMOP }, /* NOT_OP */
{ UNARY_NUMOP, ILLEGAL_NUMOP, ILLEGAL_NUMOP }, /* ISNAN_OP */
{ UNARY_NUMOP, ILLEGAL_NUMOP, ILLEGAL_NUMOP }, /* NISNAN_OP */
{ UNARY_NUMOP, UNARY_NUMOP, ILLEGAL_NUMOP }, /* INVERT_OP */
{ UNARY_NUMOP, UNARY_NUMOP, UNARY_NUMOP }, /* EXP_OP */
{ UNARY_NUMOP, UNARY_NUMOP, UNARY_NUMOP }, /* LOG_OP */
{ NARY_NUMOP, ILLEGAL_NUMOP, ILLEGAL_NUMOP }, /* MAX_OP */
{ NARY_NUMOP, ILLEGAL_NUMOP, ILLEGAL_NUMOP }, /* MIN_OP */
{ UNARY_NUMOP, ILLEGAL_NUMOP, ILLEGAL_NUMOP }, /* ABS_OP */
{ NARY_NUMOP, ILLEGAL_NUMOP, ILLEGAL_NUMOP }, /* COUNT_OP */
{ ILLEGAL_NUMOP, ILLEGAL_NUMOP, ILLEGAL_NUMOP } /* nothing */
};
/* Structure for window information */
typedef struct {
Operation operation;
Num_Operands num_operands;
int num_constants;
double constants[2];
int propagate_nan;
double illegal_value;
} Math_Data;
/* Function prototypes */
static void do_math(void *caller_data, long num_voxels,
int input_num_buffers, int input_vector_length,
double *input_data[],
int output_num_buffers, int output_vector_length,
double *output_data[],
Loop_Info *loop_info);
static void accum_math(void *caller_data, long num_voxels,
int input_num_buffers, int input_vector_length,
double *input_data[],
int output_num_buffers, int output_vector_length,
double *output_data[],
Loop_Info *loop_info);
static void start_math(void *caller_data, long num_voxels,
int output_num_buffers, int output_vector_length,
double *output_data[],
Loop_Info *loop_info);
static void end_math(void *caller_data, long num_voxels,
int output_num_buffers, int output_vector_length,
double *output_data[],
Loop_Info *loop_info);
/* Argument variables */
static int clobber = FALSE;
static int verbose = TRUE;
static int debug = FALSE;
static nc_type datatype = MI_ORIGINAL_TYPE;
static int is_signed = FALSE;
static double valid_range[2] = {0.0, 0.0};
static int copy_all_header = DEFAULT_BOOL;
static char *loop_dimension = NULL;
static int max_buffer_size_in_kb = 4 * 1024;
static double constant = DEFAULT_DBL;
static double constant2[2] = {DEFAULT_DBL, DEFAULT_DBL};
static Operation operation = UNSPECIFIED_OP;
static int propagate_nan = TRUE;
static int use_nan_for_illegal_values = TRUE;
static double value_for_illegal_operations = DEFAULT_DBL;
static int check_dim_info = TRUE;
static char *filelist = NULL;
#if MINC2
static int minc2_format = FALSE;
#endif /* MINC2 */
/* Argument table */
static ArgvInfo argTable[] = {
{NULL, ARGV_HELP, (char *) NULL, (char *) NULL,
"General options:"},
#if MINC2
{"-2", ARGV_CONSTANT, (char *) TRUE, (char *) &minc2_format,
"Produce a MINC 2.0 format output file"},
#endif /* MINC2 */
{"-clobber", ARGV_CONSTANT, (char *) TRUE, (char *) &clobber,
"Overwrite existing file."},
{"-noclobber", ARGV_CONSTANT, (char *) FALSE, (char *) &clobber,
"Don't overwrite existing file (default)."},
{"-no_clobber", ARGV_CONSTANT, (char *) FALSE, (char *) &clobber,
"Synonym for -noclobber."},
{"-verbose", ARGV_CONSTANT, (char *) TRUE, (char *) &verbose,
"Print out log messages (default)."},
{"-quiet", ARGV_CONSTANT, (char *) FALSE, (char *) &verbose,
"Do not print out log messages."},
{"-debug", ARGV_CONSTANT, (char *) TRUE, (char *) &debug,
"Print out debugging messages."},
{"-filelist", ARGV_STRING, (char *) 1, (char *) &filelist,
"Specify the name of a file containing input file names (- for stdin)."},
{"-copy_header", ARGV_CONSTANT, (char *) TRUE, (char *) ©_all_header,
"Copy all of the header from the first file."},
{"-nocopy_header", ARGV_CONSTANT, (char *) FALSE, (char *) ©_all_header,
"Do not copy all of the header from the first file."},
{"-filetype", ARGV_CONSTANT, (char *) MI_ORIGINAL_TYPE, (char *) &datatype,
"Use data type of first file (default)."},
{"-byte", ARGV_CONSTANT, (char *) NC_BYTE, (char *) &datatype,
"Write out byte data."},
{"-short", ARGV_CONSTANT, (char *) NC_SHORT, (char *) &datatype,
"Write out short integer data."},
{"-int", ARGV_CONSTANT, (char *) NC_INT, (char *) &datatype,
"Write out 32-bit integer data."},
{"-long", ARGV_CONSTANT, (char *) NC_INT, (char *) &datatype,
"Superseded by -int."},
{"-float", ARGV_CONSTANT, (char *) NC_FLOAT, (char *) &datatype,
"Write out single-precision floating-point data."},
{"-double", ARGV_CONSTANT, (char *) NC_DOUBLE, (char *) &datatype,
"Write out double-precision floating-point data."},
{"-signed", ARGV_CONSTANT, (char *) TRUE, (char *) &is_signed,
"Write signed integer data."},
{"-unsigned", ARGV_CONSTANT, (char *) FALSE, (char *) &is_signed,
"Write unsigned integer data (default if type specified)."},
{"-range", ARGV_FLOAT, (char *) 2, (char *) valid_range,
"Valid range for output data."},
{"-max_buffer_size_in_kb", ARGV_INT, (char *) 1,
(char *) &max_buffer_size_in_kb,
"Specify the maximum size of the internal buffers (in kbytes)."},
{"-dimension", ARGV_STRING, (char *) 1, (char *) &loop_dimension,
"Specify a dimension along which we wish to perform a calculation."},
{"-check_dimensions", ARGV_CONSTANT, (char *) TRUE,
(char *) &check_dim_info,
"Check that files have matching dimensions (default)."},
{"-nocheck_dimensions", ARGV_CONSTANT, (char *) FALSE,
(char *) &check_dim_info,
"Do not check that files have matching dimensions."},
{"-ignore_nan", ARGV_CONSTANT, (char *) FALSE, (char *) &propagate_nan,
"Ignore invalid data (NaN) for accumulations."},
{"-propagate_nan", ARGV_CONSTANT, (char *) TRUE, (char *) &propagate_nan,
"Invalid data in any file at a voxel produces a NaN (default)."},
{"-nan", ARGV_CONSTANT, (char *) TRUE,
(char *) &use_nan_for_illegal_values,
"Output NaN when an illegal operation is done (default)."},
{"-zero", ARGV_CONSTANT, (char *) FALSE,
(char *) &use_nan_for_illegal_values,
"Output zero when an illegal operation is done."},
{"-illegal_value", ARGV_FLOAT, (char *) 1,
(char *) &value_for_illegal_operations,
"Value to write out when an illegal operation is done."},
{NULL, ARGV_HELP, (char *) NULL, (char *) NULL,
"Options for specifying constants:"},
{"-constant", ARGV_FLOAT, (char *) 1, (char *) &constant,
"Specify a constant argument."},
{"-const", ARGV_FLOAT, (char *) 1, (char *) &constant,
"Synonym for -constant."},
{"-const2", ARGV_FLOAT, (char *) 2, (char *) constant2,
"Specify two constant arguments."},
{NULL, ARGV_HELP, (char *) NULL, (char *) NULL,
"Operations:"},
{"-add", ARGV_CONSTANT, (char *) ADD_OP, (char *) &operation,
"Add N volumes or volume + constant."},
{"-sub", ARGV_CONSTANT, (char *) SUB_OP, (char *) &operation,
"Subtract 2 volumes or volume - constant."},
{"-mult", ARGV_CONSTANT, (char *) MULT_OP, (char *) &operation,
"Multiply N volumes or volume * constant."},
{"-div", ARGV_CONSTANT, (char *) DIV_OP, (char *) &operation,
"Divide 2 volumes or volume / constant."},
{"-invert", ARGV_CONSTANT, (char *) INVERT_OP, (char *) &operation,
"Calculate 1/x at each voxel (use -constant for c/x)."},
{"-sqrt", ARGV_CONSTANT, (char *) SQRT_OP, (char *) &operation,
"Take square root of a volume."},
{"-square", ARGV_CONSTANT, (char *) SQUARE_OP, (char *) &operation,
"Take square of a volume."},
{"-abs", ARGV_CONSTANT, (char *) ABS_OP, (char *) &operation,
"Take absolute value of a volume."},
{"-max", ARGV_CONSTANT, (char *) MAX_OP, (char *) &operation,
"Synonym for -maximum."},
{"-maximum", ARGV_CONSTANT, (char *) MAX_OP, (char *) &operation,
"Find maximum of N volumes."},
{"-minimum", ARGV_CONSTANT, (char *) MIN_OP, (char *) &operation,
"Find minimum of N volumes."},
{"-exp", ARGV_CONSTANT, (char *) EXP_OP, (char *) &operation,
"Calculate c2*exp(c1*x). The constants c1 and c2 default to 1."},
{"-log", ARGV_CONSTANT, (char *) LOG_OP, (char *) &operation,
"Calculate log(x/c2)/c1. The constants c1 and c2 default to 1."},
{"-scale", ARGV_CONSTANT, (char *) SCALE_OP, (char *) &operation,
"Scale a volume: volume * c1 + c2."},
{"-clamp", ARGV_CONSTANT, (char *) CLAMP_OP, (char *) &operation,
"Clamp a volume to lie between two values."},
{"-segment", ARGV_CONSTANT, (char *) SEGMENT_OP, (char *) &operation,
"Segment a volume using range of -const2: within range = 1, outside range = 0."},
{"-nsegment", ARGV_CONSTANT, (char *) NSEGMENT_OP, (char *) &operation,
"Opposite of -segment: within range = 0, outside range = 1."},
{"-percentdiff", ARGV_CONSTANT, (char *) PERCENTDIFF_OP,
(char *) &operation,
"Percent difference between 2 volumes, thresholded (const def=0.0)."},
{"-pd", ARGV_CONSTANT, (char *) PERCENTDIFF_OP, (char *) &operation,
"Synonym for -percentdiff."},
{"-eq", ARGV_CONSTANT, (char *) EQ_OP, (char *) &operation,
"Test for integer vol1 == vol2 or vol1 == const."},
{"-ne", ARGV_CONSTANT, (char *) NE_OP, (char *) &operation,
"Test for integer vol1 != vol2 or vol1 != const."},
{"-gt", ARGV_CONSTANT, (char *) GT_OP, (char *) &operation,
"Test for vol1 > vol2 or vol1 > const."},
{"-ge", ARGV_CONSTANT, (char *) GE_OP, (char *) &operation,
"Test for vol1 >= vol2 or vol1 >= const."},
{"-lt", ARGV_CONSTANT, (char *) LT_OP, (char *) &operation,
"Test for vol1 < vol2 or vol1 < const."},
{"-le", ARGV_CONSTANT, (char *) LE_OP, (char *) &operation,
"Test for vol1 <= vol2 or vol1 <= const."},
{"-and", ARGV_CONSTANT, (char *) AND_OP, (char *) &operation,
"Calculate vol1 && vol2 (&& ...)."},
{"-or", ARGV_CONSTANT, (char *) OR_OP, (char *) &operation,
"Calculate vol1 || vol2 (|| ...)."},
{"-not", ARGV_CONSTANT, (char *) NOT_OP, (char *) &operation,
"Calculate !vol1."},
{"-isnan", ARGV_CONSTANT, (char *) ISNAN_OP, (char *) &operation,
"Test for NaN values in vol1."},
{"-nisnan", ARGV_CONSTANT, (char *) NISNAN_OP, (char *) &operation,
"Negation of -isnan."},
{"-count_valid", ARGV_CONSTANT, (char *) COUNT_OP, (char *) &operation,
"Count the number of valid values in N volumes."},
{NULL, ARGV_END, NULL, NULL, NULL}
};
/* Main program */
int main(int argc, char *argv[])
{
char **infiles, **outfiles;
int nfiles, nout;
char *arg_string;
Math_Data math_data;
Loop_Options *loop_options;
char *pname;
int num_constants;
Num_Operands num_operands;
VoxelFunction math_function;
/* Save time stamp and args */
arg_string = time_stamp(argc, argv);
/* Get arguments */
pname = argv[0];
if (ParseArgv(&argc, argv, argTable, 0) || (argc < 2)) {
(void) fprintf(stderr,
"\nUsage: %s [options] [<in1.mnc> ...] <out.mnc>\n",
pname);
(void) fprintf(stderr,
" %s -help\n\n", pname);
exit(EXIT_FAILURE);
}
nout = 1;
outfiles = &argv[argc-1];
/* Get the list of input files either from the command line or
from a file, or report an error if both are specified */
nfiles = argc - 2;
if (filelist == NULL) {
infiles = &argv[1];
}
else if (nfiles <= 0) {
infiles = read_file_names(filelist, &nfiles);
if (infiles == NULL) {
(void) fprintf(stderr,
"Error reading in file names from file \"%s\"\n",
filelist);
exit(EXIT_FAILURE);
}
}
else {
(void) fprintf(stderr,
"Do not specify both -filelist and input file names\n");
exit(EXIT_FAILURE);
}
/* Make sure that we have something to process */
if (nfiles == 0) {
(void) fprintf(stderr, "No input files specified\n");
exit(EXIT_FAILURE);
}
/* Handle special case of COUNT_OP - it always assume -ignore_nan and
-zero */
if (operation == COUNT_OP) {
propagate_nan = FALSE;
value_for_illegal_operations = 0.0;
}
/* Check that the arguments make sense */
if ((constant != DEFAULT_DBL) && (constant2[0] != DEFAULT_DBL)) {
(void) fprintf(stderr, "%s: Specify only one of -constant or -const2\n",
pname);
exit(EXIT_FAILURE);
}
if (constant != DEFAULT_DBL)
num_constants = 1;
else if (constant2[0] != DEFAULT_DBL)
num_constants = 2;
else
num_constants = 0;
num_operands = OperandTable[operation][num_constants];
if (num_operands == ILLEGAL_NUMOP) {
(void) fprintf(stderr, "%s: Operation and constants do not match.\n",
pname);
exit(EXIT_FAILURE);
}
if ((num_operands != NARY_NUMOP) && (loop_dimension != NULL)) {
(void) fprintf(stderr, "%s: Use -dimension only for cumulative ops.\n",
pname);
exit(EXIT_FAILURE);
}
if ((num_operands == UNARY_NUMOP) && (nfiles != 1)) {
(void) fprintf(stderr, "%s: Expected only one input file.\n", pname);
exit(EXIT_FAILURE);
}
if ((num_operands == BINARY_NUMOP) && (nfiles != 2)) {
(void) fprintf(stderr, "%s: Expected two input files.\n", pname);
exit(EXIT_FAILURE);
}
if ((num_operands == NARY_NUMOP) && (nfiles < 1) &&
(loop_dimension == NULL)) {
(void) fprintf(stderr, "%s: Expected at least one input files.\n",
pname);
exit(EXIT_FAILURE);
}
/* Set default copy_all_header according to number of input files */
if (copy_all_header == DEFAULT_BOOL)
copy_all_header = (nfiles == 1);
/* Set up math data structure */
math_data.operation = operation;
math_data.num_operands = num_operands;
math_data.propagate_nan = propagate_nan;
math_data.num_constants = num_constants;
switch (num_constants) {
case 1:
math_data.constants[0] = constant;
break;
case 2:
math_data.constants[0] = constant2[0];
math_data.constants[1] = constant2[1];
break;
}
if (value_for_illegal_operations != DEFAULT_DBL)
math_data.illegal_value = value_for_illegal_operations;
else if (use_nan_for_illegal_values)
math_data.illegal_value = INVALID_DATA;
else
math_data.illegal_value = 0.0;
/* Do math */
loop_options = create_loop_options();
set_loop_verbose(loop_options, verbose);
set_loop_clobber(loop_options, clobber);
#if MINC2
set_loop_v2format(loop_options, minc2_format);
#endif /* MINC2 */
set_loop_datatype(loop_options, datatype, is_signed,
valid_range[0], valid_range[1]);
if (num_operands == NARY_NUMOP) {
math_function = accum_math;
set_loop_accumulate(loop_options, TRUE, 0, start_math, end_math);
}
else {
math_function = do_math;
}
set_loop_copy_all_header(loop_options, copy_all_header);
set_loop_dimension(loop_options, loop_dimension);
set_loop_buffer_size(loop_options, (long) 1024 * max_buffer_size_in_kb);
set_loop_check_dim_info(loop_options, check_dim_info);
voxel_loop(nfiles, infiles, nout, outfiles, arg_string, loop_options,
math_function, (void *) &math_data);
free_loop_options(loop_options);
exit(EXIT_SUCCESS);
}
/* ----------------------------- MNI Header -----------------------------------
@NAME : do_math
@INPUT : Standard for voxel loop
@OUTPUT : Standard for voxel loop
@RETURNS : (nothing)
@DESCRIPTION: Routine doing math operations.
@METHOD :
@GLOBALS :
@CALLS :
@CREATED : April 25, 1995 (Peter Neelin)
@MODIFIED :
---------------------------------------------------------------------------- */
static void do_math(void *caller_data, long num_voxels,
int input_num_buffers, int input_vector_length,
double *input_data[],
int output_num_buffers, int output_vector_length,
double *output_data[],
Loop_Info *loop_info)
/* ARGSUSED */
{
Math_Data *math_data;
long ivox;
double value1, value2;
double illegal_value;
Operation operation;
int num_constants, iconst;
double constants[2];
/* Get pointer to window info */
math_data = (Math_Data *) caller_data;
/* Check arguments */
if ((input_num_buffers > 2) || (output_num_buffers != 1) ||
(output_vector_length != input_vector_length)) {
(void) fprintf(stderr, "Bad arguments to do_math!\n");
exit(EXIT_FAILURE);
}
/* Get info */
operation = math_data->operation;
num_constants = math_data->num_constants;
for (iconst=0; iconst < sizeof(constants)/sizeof(constants[0]); iconst++) {
if (iconst < num_constants)
constants[iconst] = math_data->constants[iconst];
else if ((operation == INVERT_OP) ||
(operation == EXP_OP) ||
(operation == LOG_OP))
constants[iconst] = 1.0;
else
constants[iconst] = 0.0;
}
illegal_value = math_data->illegal_value;
/* Set default second value */
value2 = constants[0];
/* Loop through the voxels */
for (ivox=0; ivox < num_voxels*input_vector_length; ivox++) {
value1 = input_data[0][ivox];
if (input_num_buffers == 2)
value2 = input_data[1][ivox];
if ((value1 == INVALID_DATA) || (value2 == INVALID_DATA)) {
switch(operation) {
case ISNAN_OP:
output_data[0][ivox] = 1.0;
break;
case NISNAN_OP:
output_data[0][ivox] = 0.0;
break;
default:
output_data[0][ivox] = INVALID_DATA;
break;
}
}
else {
switch (operation) {
case ADD_OP:
output_data[0][ivox] = value1 + value2; break;
case SUB_OP:
output_data[0][ivox] = value1 - value2; break;
case MULT_OP:
output_data[0][ivox] = value1 * value2; break;
case DIV_OP:
if (value2 != 0.0)
output_data[0][ivox] = value1 / value2;
else
output_data[0][ivox] = illegal_value;
break;
case INVERT_OP:
if (value1 == 0.0)
output_data[0][ivox] = illegal_value;
else
output_data[0][ivox] = value2 / value1;
break;
case SQRT_OP:
if (value1 < 0.0)
output_data[0][ivox] = illegal_value;
else
output_data[0][ivox] = sqrt(value1);
break;
case SQUARE_OP:
output_data[0][ivox] = value1 * value1; break;
case ABS_OP:
if (value1 < 0.0)
output_data[0][ivox] = -value1;
else
output_data[0][ivox] = value1;
break;
case EXP_OP:
output_data[0][ivox] = constants[1] * exp(value1 * constants[0]);
break;
case LOG_OP:
if ((value1 <= 0.0) || (constants[1] <= 0.0) ||
(constants[0] == 0.0))
output_data[0][ivox] = illegal_value;
else
output_data[0][ivox] = log(value1/constants[1])/constants[0];
break;
case SCALE_OP:
output_data[0][ivox] = value1 * constants[0] + constants[1]; break;
case CLAMP_OP:
if (value1 < constants[0])
value1 = constants[0];
else if (value1 > constants[1])
value1 = constants[1];
output_data[0][ivox] = value1;
break;
case SEGMENT_OP:
if ((value1 < constants[0]) || (value1 > constants[1]))
output_data[0][ivox] = 0.0;
else
output_data[0][ivox] = 1.0;
break;
case NSEGMENT_OP:
if ((value1 < constants[0]) || (value1 > constants[1]))
output_data[0][ivox] = 1.0;
else
output_data[0][ivox] = 0.0;
break;
case PERCENTDIFF_OP:
if ((value1 < constants[0]) || (value1 == 0.0))
output_data[0][ivox] = illegal_value;
else {
output_data[0][ivox] = 100.0 * (value1 - value2) / value1;
}
break;
case EQ_OP:
output_data[0][ivox] = (((rint(value1)-rint(value2)) == 0.0)
? 1.0 : 0.0);
break;
case NE_OP:
output_data[0][ivox] = (((rint(value1)-rint(value2)) != 0.0)
? 1.0 : 0.0);
break;
case GT_OP:
output_data[0][ivox] = value1 > value2; break;
case GE_OP:
output_data[0][ivox] = value1 >= value2; break;
case LT_OP:
output_data[0][ivox] = value1 < value2; break;
case LE_OP:
output_data[0][ivox] = value1 <= value2; break;
case AND_OP:
output_data[0][ivox] =
(((rint(value1) != 0.0) && (rint(value2) != 0.0)) ? 1.0 : 0.0);
break;
case OR_OP:
output_data[0][ivox] =
(((rint(value1) != 0.0) || (rint(value2) != 0.0)) ? 1.0 : 0.0);
break;
case NOT_OP:
output_data[0][ivox] = ((rint(value1) == 0.0) ? 1.0 : 0.0);
break;
case ISNAN_OP:
output_data[0][ivox] = 0.0; /* To get here, value is not nan */
break;
case NISNAN_OP:
output_data[0][ivox] = 1.0;
break;
default:
(void) fprintf(stderr, "Bad op in do_math!\n");
exit(EXIT_FAILURE);
}
}
}
return;
}
/* ----------------------------- MNI Header -----------------------------------
@NAME : accum_math
@INPUT : Standard for voxel loop
@OUTPUT : Standard for voxel loop
@RETURNS : (nothing)
@DESCRIPTION: Routine for doing accumulation math operations.
@METHOD :
@GLOBALS :
@CALLS :
@CREATED : April 25, 1995 (Peter Neelin)
@MODIFIED :
---------------------------------------------------------------------------- */
static void accum_math(void *caller_data, long num_voxels,
int input_num_buffers, int input_vector_length,
double *input_data[],
int output_num_buffers, int output_vector_length,
double *output_data[],
Loop_Info *loop_info)
/* ARGSUSED */
{
Math_Data *math_data;
long ivox;
double value, oldvalue;
Operation operation;
int propagate_nan;
/* Get pointer to window info */
math_data = (Math_Data *) caller_data;
/* Check arguments */
if ((input_num_buffers != 1) || (output_num_buffers != 1) ||
(output_vector_length != input_vector_length)) {
(void) fprintf(stderr, "Bad arguments to accum_math!\n");
exit(EXIT_FAILURE);
}
/* Get info */
operation = math_data->operation;
propagate_nan = math_data->propagate_nan;
/* Loop through the voxels */
for (ivox=0; ivox < num_voxels*input_vector_length; ivox++) {
/* Get previous value and the next value */
oldvalue = output_data[0][ivox];
value = input_data[0][ivox];
/* If the new data is invalid, then either mark the output as invalid
or ignore it */
if (value == INVALID_DATA) {
if (propagate_nan) {
output_data[0][ivox] = INVALID_DATA;
}
}
/* If we haven't set anything yet, then just copy the new value */
else if (oldvalue == UNINITIALIZED_DATA) {
output_data[0][ivox] = input_data[0][ivox];
}
/* Do the operation if the old data and the new data are valid */
else if (oldvalue != INVALID_DATA) {
switch (operation) {
case ADD_OP:
output_data[0][ivox] = oldvalue + value;
break;
case MULT_OP:
output_data[0][ivox] = oldvalue * value;
break;
case AND_OP:
output_data[0][ivox] =
(((oldvalue != 0.0) && (rint(value) != 0.0)) ? 1.0 : 0.0);
break;
case OR_OP:
output_data[0][ivox] =
(((oldvalue != 0.0) || (rint(value) != 0.0)) ? 1.0 : 0.0);
break;
case MAX_OP:
if (value > oldvalue)
output_data[0][ivox] = value;
break;
case MIN_OP:
if (value < oldvalue)
output_data[0][ivox] = value;
break;
case COUNT_OP:
output_data[0][ivox]++;
break;
default:
(void) fprintf(stderr, "Bad op in accum_math!\n");
exit(EXIT_FAILURE);
}
}
} /* Loop over voxels */
return;
}
/* ----------------------------- MNI Header -----------------------------------
@NAME : start_math
@INPUT : Standard for voxel loop
@OUTPUT : Standard for voxel loop
@RETURNS : (nothing)
@DESCRIPTION: Start routine for math accumulation.
@METHOD :
@GLOBALS :
@CALLS :
@CREATED : April 25, 1995 (Peter Neelin)
@MODIFIED :
---------------------------------------------------------------------------- */
static void start_math(void *caller_data, long num_voxels,
int output_num_buffers, int output_vector_length,
double *output_data[],
Loop_Info *loop_info)
/* ARGSUSED */
{
Math_Data *math_data;
long ivox;
/* Get pointer to window info */
math_data = (Math_Data *) caller_data;
/* Check arguments */
if (output_num_buffers != 1) {
(void) fprintf(stderr, "Bad arguments to start_math!\n");
exit(EXIT_FAILURE);
}
/* Get info */
operation = math_data->operation;
/* Loop through the voxels, marking them all as uninitialized. We treat
COUNT_OP as a special case since it always has a value. This is
especially important to prevent it from going through
the code in accum_math for handling the first valid voxel which
just assigns the first value. */
for (ivox=0; ivox < num_voxels*output_vector_length; ivox++) {
switch (operation) {
case COUNT_OP:
output_data[0][ivox] = 0.0;
break;
default:
output_data[0][ivox] = UNINITIALIZED_DATA;
break;
}
}
return;
}
/* ----------------------------- MNI Header -----------------------------------
@NAME : end_math
@INPUT : Standard for voxel loop
@OUTPUT : Standard for voxel loop
@RETURNS : (nothing)
@DESCRIPTION: Start routine for math accumulation.
@METHOD :
@GLOBALS :
@CALLS :
@CREATED : April 25, 1995 (Peter Neelin)
@MODIFIED :
---------------------------------------------------------------------------- */
static void end_math(void *caller_data, long num_voxels,
int output_num_buffers, int output_vector_length,
double *output_data[],
Loop_Info *loop_info)
/* ARGSUSED */
{
Math_Data *math_data;
long ivox;
double value;
double illegal_value;
/* Get pointer to window info */
math_data = (Math_Data *) caller_data;
/* Check arguments */
if (output_num_buffers != 1) {
(void) fprintf(stderr, "Bad arguments to end_math!\n");
exit(EXIT_FAILURE);
}
/* Get info */
operation = math_data->operation;
illegal_value = math_data->illegal_value;
/* Loop through the voxels, checking for uninitialized values */
for (ivox=0; ivox < num_voxels*output_vector_length; ivox++) {
value = output_data[0][ivox];
if ((value == UNINITIALIZED_DATA) || (value == INVALID_DATA)) {
output_data[0][ivox] = illegal_value;
}
}
return;
}
|