imresize incorrectly clips output to [0,1]

Stefan Gustavson stefan.gustavson at gmail.com
Fri Apr 3 08:08:30 CDT 2009


Bug report for Octave 3.0.3 configured for i686-pc-msdosmsvc

Description:
-----------

The function imresize(img, [w h], interp), when using double input,
incorrectly clips its output to the range [0,1]. This is both wrong and
inconsistent with what Matlab does. Many image processing applications
use a wider range of values than [0,1]. Negative values are only to be
disallowed in images representing pictures, and even those have no
inherent upper limit. Clipping to a certain range is an issue with image
file formats with a limited precision and range (e.g. 8-bit samples),
not with the image data as such.

If arbitrary clipping to [0,1] is performed for any other image
processing functions, they need to be fixed as well. I am sorry,
but I have not found the time to check the entire "image" package.

Repeat-By:
---------

img = 10*(rand([64, 64])-0.5); % Range of img is [-5, 5]
img2 = imresize(img, [64, 64]); % Essentially a no-op
% Range of img2 is clipped to [0,1], as demonstrated by:
min(img2(:))
max(img2(:))

Fix:
---

Remove the three lines marked with "->" from "imremap.m":

  ## Change the class of the results according to the class of the image
  c = class(im);
  if (strcmpi(c, "double"))
->    ## Remove overshooting caused by bicubic interpolation
->   warped(warped>1) = 1;
->   warped(warped<0) = 0;
  elseif (strcmpi(c, "uint8"))
    warped = uint8(warped);
  elseif (strcmpi(c, "uint16"))
    warped = uint16(warped);
  endif


Configuration (please do not edit this section):
-----------------------------------------------

uname output: Windows
configure opts:
Fortran compiler: fc-msvc
FFLAGS: -O2
F2C: @F2C@
F2CFLAGS: @F2CFLAGS@
FLIBS: -lhdf5 -lzlib -lf2c -lkernel32
CPPFLAGS: -I. -Ic:/Software/VCLibs/include
INCFLAGS: -I. -I. -I./liboctave -I./src -I./libcruft/misc
C compiler: cc-msvc, version
CFLAGS: -O2 -MD
CPICFLAG:
C++ compiler: cc-msvc, version
CXXFLAGS: -O2 -EHs -MD
CXXPICFLAG:
LD_CXX: cc-msvc
LDFLAGS:
LIBFLAGS: -L.
RLD_FLAG:
BLAS_LIBS: -llapack -lblas
FFTW_LIBS: -lfftw3
LIBS: -lreadline -lncurses -lhdf5 -lzlib -lws2_32 -lkernel32
LEXLIB:
LIBGLOB: -lglob
SED: /usr/bin/sed
DEFS:
-DPACKAGE_NAME="" -DPACKAGE_TARNAME="" -DPACKAGE_VERSION=""
-DPACKAGE_STRING="" -DPACKAGE_BUGREPORT="" -DOCTAVE_SOURCE=1
-D_GNU_SOURCE=1 -DSTDC_HEADERS=1 -DHAVE_SYS_TYPES_H=1
-DHAVE_SYS_STAT_H=1 -DHAVE_STDLIB_H=1 -DHAVE_STRING_H=1
-DHAVE_MEMORY_H=1 -DHAVE_INTTYPES_H=1 -DHAVE_STDINT_H=1
-DHAVE_UNISTD_H=1 -DSEPCHAR=';' -DSEPCHAR_STR=";"
-D__NO_MATH_INLINES=1 -DCXX_NEW_FRIEND_TEMPLATE_DECL=1
-DCXX_ISO_COMPLIANT_LIBRARY=1 -DCXX_ABI=unknown -DHAVE_QHULL=1
-DHAVE_PCRE=1 -DHAVE_ZLIB_H=1 -DHAVE_ZLIB=1 -DHAVE_HDF5_H=1
-DHAVE_HDF5=1 -DHAVE_H5GGET_NUM_OBJS=1 -D_HDF5USEDLL_=1 -DHAVE_FFTW3=1
-DHAVE_GLPK_H=1 -DHAVE_GLPK=1 -DHAVE_CURL_CURL_H=1 -DHAVE_CURL=1
-DHAVE_IEEE754_DATA_FORMAT=1 -DF77_FUNC(name,NAME)=name ## _
-DF77_FUNC_(name,NAME)=name ## __ -DHAVE_BLAS=1
-DHAVE_SUITESPARSE_UMFPACK_H=1 -DHAVE_UMFPACK=1
-DUMFPACK_SEPARATE_SPLIT=1 -DHAVE_SUITESPARSE_COLAMD_H=1
-DHAVE_COLAMD=1 -DHAVE_SUITESPARSE_CCOLAMD_H=1 -DHAVE_CCOLAMD=1
-DHAVE_SUITESPARSE_CHOLMOD_H=1 -DHAVE_CHOLMOD=1
-DHAVE_SUITESPARSE_CS_H=1 -DHAVE_CXSPARSE=1 -Dmode_t=int -Dpid_t=int
-Duid_t=int -Dgid_t=int -DHAVE_DEV_T=1 -DHAVE_INO_T=1
-DHAVE_LONG_LONG_INT=1 -DHAVE_UNSIGNED_LONG_LONG_INT=1
-DHAVE_SIG_ATOMIC_T=1 -DSIZEOF_SHORT=2 -DSIZEOF_INT=4 -DSIZEOF_LONG=4
-DSIZEOF_LONG_LONG=8 -DHAVE_ALLOCA=1 -DNPOS=std::string::npos
-DHAVE_PLACEMENT_DELETE=1 -DSTDC_HEADERS=1 -DHAVE_ASSERT_H=1
-DHAVE_DIRECT_H=1 -DHAVE_FCNTL_H=1 -DHAVE_FLOAT_H=1
-DHAVE_INTTYPES_H=1 -DHAVE_LIMITS_H=1 -DHAVE_LOCALE_H=1
-DHAVE_MEMORY_H=1 -DHAVE_STDINT_H=1 -DHAVE_STDLIB_H=1
-DHAVE_STRING_H=1 -DHAVE_SYS_STAT_H=1 -DHAVE_SYS_TYPES_H=1
-DHAVE_SYS_UTIME_H=1 -DHAVE_UNISTD_H=1 -DHAVE_VARARGS_H=1
-DHAVE_SSTREAM=1 -DHAVE_GLOB_H=1 -DHAVE_FNMATCH_H=1 -DHAVE_CONIO_H=1
-DHAVE_ATEXIT=1 -DHAVE_CHMOD=1 -DHAVE_DUP2=1 -DHAVE_EXECVP=1
-DHAVE_GETCWD=1 -DHAVE_GETPID=1 -DHAVE__KBHIT=1 -DHAVE_MEMMOVE=1
-DHAVE_MKDIR=1 -DHAVE_PUTENV=1 -DHAVE_RAISE=1 -DHAVE_RENAME=1
-DHAVE_RMDIR=1 -DHAVE_SETLOCALE=1 -DHAVE_SETVBUF=1 -DHAVE_STAT=1
-DHAVE_STRDUP=1 -DHAVE_STRERROR=1 -DHAVE_STRICMP=1 -DHAVE_STRNICMP=1
-DHAVE_TEMPNAM=1 -DHAVE_UMASK=1 -DHAVE_UNLINK=1 -DHAVE_UTIME=1
-DHAVE_VFPRINTF=1 -DHAVE_VSPRINTF=1 -DHAVE_VSNPRINTF=1 -DHAVE__CHMOD=1
-DHAVE__SNPRINTF=1 -DHAVE__UTIME32=1 -DOCTAVE_HAVE_BROKEN_STRPTIME=1
-D_USE_MATH_DEFINES=1 -DHAVE_LOADLIBRARY_API=1
-DENABLE_DYNAMIC_LINKING=1 -DHAVE__FINITE=1 -DHAVE__ISNAN=1
-DHAVE__COPYSIGN=1 -DHAVE_DECL_SIGNBIT=0 -DHAVE_STRUCT_STAT_ST_RDEV=1
-DHAVE_DECL_TZNAME=1 -DHAVE_TZNAME=1 -DCLOSEDIR_VOID=1
-DMKDIR_TAKES_ONE_ARG=1 -DUSE_READLINE=1 -DRETSIGTYPE=void
-DHAVE_DECL_SYS_SIGLIST=0 -DMUST_REINSTALL_SIGHANDLERS=1
-DRETSIGTYPE_IS_VOID=1 -DGNUPLOT_BINARY="pgnuplot"


User-preferences (please do not edit this section):

EDITOR = C:\Program Files\Octave\tools\wscite\SciTE.exe
EXEC_PATH = C:\Program Files\Octave\msys\bin;C:\Program
Files\Octave\libexec\octave\3.0.3\site\exec\i686-pc-msdosmsvc;C:\Program
Files\Octave\libexec\octave\api-v32\site\exec\i686-pc-msdosmsvc;C:\Program
Files\Octave\libexec\octave\site\exec\i686-pc-msdosmsvc;C:\Program
Files\Octave\libexec\octave\3.0.3\exec\i686-pc-msdosmsvc;C:\Program
Files\Octave\bin;C:\Program Files\Octave;C:\Program
Files\Aqsis\bin;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\Program
Files\ATI Technologies\ATI.ACE\Core-Static;C:\Program
Files\MATLAB\R2007b\bin;C:\Program Files\MATLAB\R2007b\bin\win32
IMAGE_PATH = .;C:\Program Files\Octave\share\octave\3.0.3\imagelib
PAGER = less
PS1 = \s:\#>
PS2 = >
PS4 = +
beep_on_error = 0
completion_append_char =
crash_dumps_octave_core = 1
echo_executing_commands = 0
fixed_point_format = 0
gnuplot_binary = pgnuplot
gnuplot_command_end =

gnuplot_command_plot = pl
gnuplot_command_replot = rep
gnuplot_command_splot = sp
gnuplot_command_title = t
gnuplot_command_using = u
gnuplot_command_with = w
history_file = C:\Documents and Settings\stegu\.octave_hist
history_size = 1024
ignore_function_time_stamp = system
info_file = C:\Program Files\Octave\share\info\octave.info
info_program = info
makeinfo_program = makeinfo
max_recursion_depth = 256
output_max_field_width = 5
output_precision = 5
page_output_immediately = 0
page_screen_output = 1
print_answer_id_name = 1
print_empty_dimensions = 1
save_precision = 16
saving_history = 1
sighup_dumps_octave_core = 1
sigterm_dumps_octave_core = 1
silent_functions = 0
split_long_rows = 1
string_fill_char =
struct_levels_to_print = 2
suppress_verbose_help_message = 0


More information about the Bug-octave mailing list