File Release Notes and Changelog
Notes:
AMIDE is a tool for viewing and analyzing medical image data sets.
It's capabilities include the simultaneous handling of multiple data
sets imported from a variety of file formats, image fusion, 3D region
of interest drawing and analysis, volume rendering, and rigid body
alignments.
Changes:
0.9.1 2007.12.16
* added gtk_disable_locale on win32 to keep i18n from screwing up
* import raw on win32 was crashing due to NULL string [1]
* transformed help docs into gnome-doc-utils format
* fixed .spec file, rpm build now working again
[1] Found by Kai-Hsiang Chuang
0.9.0 2007.12.02
* removing gtk 2.2 dependencies, now dependent on gtk >= 2.10
-gtk_option_menu switched to gtk_combo
-gnome_config use switched to GConf on unix
-config settings now saved in registry on win32
-bug with dragging of off canvases now gone
* removed dependencies on libgnome and libgnomeui
-GNOMEUI menu/toolbar switched to GtkAction
-GnomeDruid replaced with GtkAssistant
-gnome_program_init replaced with gtk_init
* amitk_space_transform now propagates correctly through children
* xpm icons removed, all icons now from .h pixbuf data
* dicom import now reads in full series date and time [1]
* now can export data sets via libdcmdata [2,3]
* fixed errors with cropping type conversion
* change Categories in .desktop file [4]
* extended math capabilities for multi-frame data sets [5]
* changed labeling of %ID/g to %ID/cc to make explicit [6]
[1] Requested by David Yerushalmi
[2] Requested by Aurlie Dutour
[3] Requested by Matthew Leonard
[4] Change by Dominique Belhachemi
[5] Requested by Michael Braden
[6] Requested by James Yu
0.8.23
* help window incorrect for freehand roi's [1]
* changed erase outside roi to shift-m3 from shift-ctrl-m3
* incorrect exporting of multi-frame data sets, values zero [2]
* poor threshold setting on all negative value data sets [3]
* import file dialog now closed prior to importing
[1] Found by Marco Bucci
[2] Found by Andre Manook
[3] Noticed by David Yerushalmi
0.8.22 2007.09.28
* fixed bug with drawing 1st voxel of freehand (introduced 0.8.21) [1]
* autodetect xmedcon format support (requires xmedcon > 0.9.10)
* button to set the view center as origin [2]
[1] Found by Jeff Arkles
[2] Requested by Michael Braden
0.8.21 2007.07.14
* switched index to strchr in xml.c, amitk_point.c
* man page fix [1]
* freehand roi's now initially line up with the voxels of the dataset
* hopefully finally now have a true radix sign fix, as mingw can't
really use setlocale
* flip around NIFTI images to get right orientation [2]
[1] From Dominique Belhachemi
[2] Suggested by Buyean Lee
0.8.20 2007.06.29
* message popup for export raw data dimensions
* init allocated density memory to 0 in render.c
* check that voxel sizes are valid in imported data sets [1]
* tightened up raw_data_import dialog
* added millimeter indicators to crop dialog [2]
* update for NIFTI support in XMedCon [3]
* even more correct placement/orientation of dicom data [3]
* added ability to output PNG instead of JPG [4]
* added a contour form of hot metal [4]
* add a preference for the initial directory to look in [4]
* added ability to manually enter in the rotation transform [4]
* fixed a bug with rotations
* internationalization, now uses en_US instead of POSIX on windows,
also guesses if incorrect radix sign, and tries to fix
[1] Suggested by Gokhan Ozer
[2] Suggested by Luca Guerrieri
[3] Suggested by Buyean Lee
[4] Suggested by Andre Manook
0.8.19 2006.10.22
* fixed string error in dcmtk_interface.cc [1]
* somewhat more correct placement/orientation of dicom data
* fixed bug for reading in dynamic dicom data
* DCMTK returns number strings as POSIX [2]
* added progress bars for filters
* object movement in alignment wizard now propagates correctly
[1] Found by Kris Thielemans
[2] Found by Marco Bucci
0.8.18 2006.10.08
* validate strings from xmedcon as utf8 [1]
* another potential fix for 64bit check xml_check_file_32bit_okay [2]
* added export dataset dialog
- allows selection of output voxel size [3]
- allows selecting bounding box (tight or inclusive) [3]
* better guessing for reading in DICOM datasets [4,5]
* first try at DICOM gated support [5]
* added ability to import objects (like ROI's) from previous .xif's [6]
* better handling of edges in tri-linear interpolation [7]
* active data set now required to be visible when moved/rotated
[1] Found by Buyean Lee
[2] Found by Marco Bucci
[3] Suggested by Luca Guerrieri
[4] Help from Han Chunlei
[5] Help from Kris Thielemans
[6] Suggested by Marco Bucci and many others
[7] Suggested by Patrick Chow
0.8.17 2006.08.30
* added toolbox for doing image math between data sets [1]
* now saves weight/dob/dose/etc. when exporting with xmedcon [1]
* fixed bug loading DICOM on windows [2]
[1] with help from Matthew Leonard
[2] found by Martin Rodriguez-Porcel
0.8.16 2006.06.29
* wasn't reading in intercepts correctly from .xif files [1]
* potential fix for 64 bit check in xml_check_file_32bit_okay [2]
* fixed bug with rendering of ROI's [3]
* in help panel, button "1" now refered to as "m1" for clarity
[1] Found by Peter Bloomfield and Marwan Elkoussy
[2] Found by Jose Manuel Quesada
[3] Found by John Storey-Bishoff
0.8.15.3 2006.05
* export->series for dynamic data sets led to core dump [1]
* now handles DICOM files without scale factors (eg MR)
[1] Found by Tove Olafsen
0.8.15.2 2006.04.15
* fix to libecat support - includes calibration when possible
* fixed 64bit/32bit file detection on Mac OS X
* fixed several glitches with DICOM handling [1]
- better guestimation of slice thickness
- correctly handles patient orientation
[1] Found by Andre Manook
0.8.15.1 2006.04.09
* slight xmedcon tweaks - now require xmedcon >= 0.9.9.0
* various small fixes for gcc 4.1 (fedora core 5)
* fixed independent threshold - menu now gets enabled properly
* modified libmdc/libecat support, allow reading in truncated data sets
* fix for exporting data sets
0.8.15 2006.02.21
* added independent color tables for 2nd and 3rd sets of views [1]
* strip NaN's on export->resliced data set to make xmedcon happy
* added internal scaling intercept - more compatible with dicom
- intercept gets removed if all are zero
* added additional support for DICOM through the DCMTK library
- has support for reading in DICOMDIR files
* updates for MinGW build
* added a field of view parameter - to look at only a subset of data
* now reduces scale factor dimension of imported data when possible
* can now specify how multiple canvases are layed out (panel_layout option)
[1] Suggested by Freimut Juengling
[2] Suggested by Kris Thielemans
0.8.14 2005.11.17
* faster functions for initialized amitk_raw_data structures [1]
* fast roi quantitation now gives more accurate median value
* freehand/isocontour drawing speed-ups
* can export underlying raw roi values to file [2]
* canvases and tree/help can torn off as separate windows [3]
* removed 'logic' for placing orthogonal sets of canvases
-now simply left to right
* Default rendering opacity can now be specified as density only [4]
* raw data import dialog values are now persistent [5]
* distribution on threshold now only shown for global thresholding
* can now display thresholds as center/window as well as min/max [6]
* additional CT windowing options [6]
* added "subject orientation", and appropriate displays on canvas [6]
* added subject name, id, and dob entries [6]
* toolbox dialogs now correctly free memory on destroy signal
[1] Requested by Buyean Lee
[2] Requested by Daniel Rubin
[3] Suggested by Jonathan Bailleul
[4] Suggested by Jeff Tseng
[5] Suggested by Robert Hunter
[6] Suggested by Freimut Juengling
0.8.13 2005.06.24
* added freehand ROI's [1]
* added drawing mode for isocontour ROI's
* Makefile update for gnome 2.10 [2]
[1] Requested by Buyean Lee and many other people
[2] Found by CG Hsu
0.8.12.1 2005.06.16
* slight tweek to the zoom setting code
0.8.12 2005.06.15
* added an option to recover corrupted xif files
* various small bug fixes for some warnings generated by gcc 4.0
* changing voxel_dim (from adding or removing data sets) now compensates zoom [1]
* analysis now uses NAN for undefined stuff if available
* slice functions now return NAN for undefined pixels instead of zero
* implemented function to output multiple data sets as single file [2]
* can now put a time label on flythrough and rendered movies [3]
* new dynamic movie type - frame with smoothing
* three isocontour ranges - above min, below max, between min/max [4]
[1] Found by Steve Jones
[2] Requested by Sheena Lee
[3] Requested by Anna Wu
[4] Requested by Karen Herbst
0.8.11 2005.03.12
* ROI analysis wasn't correctly freeing the pointer array [1]
* disabled reporting of "total" statistic, as this is often
wrong depending on the underlying data type
* error in saving of .xif files with dynamic data on win32 [2]
[1] Found by Michael Kreissl
[2] Found by Steve Jones
0.8.10 2005.01.26
* ui_series no longer requires corresponding ui_study canvas to be shown [1]
* Slices of thickness greater then the voxel size were offset in the depth direction [2]
[1] Found by Peter Bloomfield
[2] Found by Jeff Tseng
0.8.9.1 2005.01.13
* another try at a win32 internationalization fix
[1] Found by Alien and Liesje
0.8.9 2005.01.08
* Makefile.am problems for compiling w/o xmedcon [1]
* saving roi statistics records isocontour value [2]
* ui_study titlebar includes filename as well as study name[2]
* internationalization problems [3,4]
* set default action on dialogs [3]
* clean building without scrollkeeper [3]
* center of an isocontour now defined as the center of mass [5]
[1] Found by Iain
[2] Requested by Steve Jones
[3] Found/Requested by Alien
[4] Found by Erik Nolf
[5] Requested by Steve Jones
0.8.8.1 2004.11.10
* updates for x86_64 compilation
0.8.8 2004.11.02
* crash when poping up object dialog of non-viewed data set [1]
* accurate roi quantitation algorithm didn't work in inverse mode [2]
* isocontour's were overdrawing by one voxel [3]
* can now add roi to study object in tree
* new preference: can have isocontours drawn as filled
* incorrect use of MdcSwapBytes was causing rotated images on big endian [4]
* added ability to specify isocontour value (lower value) [3]
* updates for (X)MedCon 0.9.8.0
[1] Found by Patrick Chow
[2] Found by David Stout
[3] Found/Requested by Steve Jones
[4] Found by Anna Wu
0.8.7 2004.08.22
* spin buttons on roi initial popup didn't take decimal places [1]
* ui_series panel size can now be adjusted [2]
* new isocontour, wasn't showing image value on help window [3]
* fixed isocontour erasing [3]
* fixed error in profile tool, incorrect values for view_duration < frame_duration [4]
* users now have the option of a more accurate roi quantitation algorithm [5]
* increased accuracy of the fast roi quantitation algorithm
[1] Found by Keon Kang
[2] Requested by Freimut Juengling
[3] Found by Steve Jones
[4] Found by Patrick Chow
[5] Found by Peter Bloomfield
0.8.6 2004.06.22
* raw data import dialog didn't take gates into account for calculating bytes
* fixed rendered movie generation
* can now calculate inside ROI for values greater than specific value
* amitk_xif_sel crashed when browsing "/" directory [1]
* fixed importing of analyze files [2]
* get configure script to end if m4 not found [3]
* rpm now builds correctly on FC2
[1] Found by Peter Bloomfield
[2] Found by Charles Henri Malbert
[3] Found by Emily Heath
0.8.5 2004.06.07
* fixed fopen binary versus text bug on windows [1]
* export view function now includes roi's, etc. [1]
* fly through mpeg's now include roi's, etc.
* data sets can now also be exported resliced [2]
* slice generation slightly more accurate for iterating over depth
* fixed core dumps with filtering [3]
* added export jpeg abilities to ui_series
* added gated study support to ui_series
* added ability to select which objects go onto series on init
* added ability to select which objects go into rendering on init
* libecat strings aren't necesarrily NULL terminated [4]
* some gtk 2.4 updates (switch to g_idle)
* switched to libmdc progressive loader
* values read through libmdc are now quantified & calibrated,
they had only been quantified before (branching factor hadn't been added)
* updates to gtk-doc reference doc generation process
[1] Found/Requested by Steve Jones
[2] Requested by Jayashree Kanchana
[3] Found by Vania Kenanova
[4] Found by Kris Thielemans
0.8.4 2004.04.26
* fixed bug in *_calc_frame_max_min [1]
[1] Found by Peter Bloomfield
0.8.3 2004.04.21
* added support for gated files [1,3]
* profile tool now updated automatically on data set (de)select
* fixed small graphical glitch with line between full/scaled bars
* fixed errors with scale_factor and filtering/cropping [1]
* fixed crash with double clicking finish on filter/crop
* can now validly pop up a dialog on an undrawn roi [2]
* added ROI quantitation based on Lee et al, Nuc Med Comm 2000 21:685-690
[1] Found/Suggested by Chris Deroose
[2] Found/Suggested by Keon Kang
[3] Suggested by David Stout
0.8.2 2004.02.16
* incorrect loading of frame duration through libecat [1]
* rendering conversion had diff. dimension between slice and rendering
lead to core dump on Mac OS X, and with some gcc's [2, 5]
* SUV calculations were off by a factor of 1000 [3]
* ignore gdk warnings on win32
work around for gdkproperty.c "Not Implemented" warnings (gtk 2.2.4)
* got popt/icons working on win32
* updated win32 install. .xif files now associated with AMIDE
* updated most of the icons (exception of the toolbar icons)
* better interactivity on threshold for remote displays [1]
* roi analysis "copy to clipboard" now left justifies roi names [3]
* max/min calculations now done on demand
* changed mechanism for changing isocontours
old mechanism wasn't working on win32
* added x limits to the gaussian fit on the profile tool [1]
* added scientific notation support to most of the spin buttons [2]
[1] Found/Suggested by Kris Thielemans
[2] Found by Anna Wu
[3] Found/Suggested by Chris Deroose
[4] Found by Keon Kang
[5] Found by Chris Heier
0.8.1 2004.01.13
* specifying import types was messed up [1]
[1] Found by Patrick Chow & Peter Bloomfield
0.8.0 2004.01.11
* finally added a (much requested) profile tool
* improved (correct) loading of non-float data files via (X)MedCon
* can drag file's from Nautilus onto the tree
* updated for new dynamic data structure coming in (X)MedCon 0.9.0
* preferences now saved on a per study/data set basis
default preferences used only for new studies/data sets
* added the ability to set a bone and soft tissue window for CT data sets
* removed standard error from roi stats. wasn't sure if it was calculated
correctly (wasn't weighted), and it's not the experimental std. err. anyway
* added copy button to roi statistics window (for copy and paste of values)
* various AmitkThreshold cleanups
* saving on windows to samba share crashed for large files [1]
not sure why... work around is saving in 16MB chunks in amitk_raw_data.c
* default roi stat file is probably annotated as .tsv (was .csv)
* warning messages to console/dialogs is now a preference, (was compiled in)
* added ability to export data sets
* i18n broke reading in .xif's, concorde files for European locales [2]
* was trying to change permissions on non-existant files [3]
* reading from libecat7 versus libmdc was setting offset inconsistently [3]
* filter fwhm was off by a factor of 2
* roi analysis results now have more precision, as does the time dialog
* Spanish translation of the manual [4]
* got rid of console on windows
[1] Found by Patrick Chow
[2] Found by Chris Deroose
[3] Found by Kris Thielemans
[4] Contributed by Pablo Sao
0.7.14 2003.08.15
* disabled anti-aliasing on canvas, was getting core dumps on ROI resizing.
* problems with self-referencing with slice_parent fixed, data sets now freed
* crop wizard now allows data type/scaling conversions too
* updates to help in rois.xml
* added help button to several dialogs
* fixed small bug with rendering isocontours alone
* removed data type specific crop and projection functions
complexity not worth incremental performance improvement
* amitk_data_set_get_projection now does all 3 planes at once
* cursor color now changes immediately when data set changed
* erase outside roi no longer core dumps
* when voxel sizes of a data set are changed, the children
now try to follow the scaling changes.
0.7.13 2003.07.09
* fixed problems with selecting XIF files in the file selector
0.7.12 2003.07.05
* added ability to save XIF files as flat file, made default
* put in configure ability to detect libgnomecanvas version,
disable anti-aliases canvases previous to libgnomecanvas 2.2.0
* .iss file now configure time generated
* configure.in file tweaks
* zh_CN i18n update from Wales Wang
* threshold "abolute" legend now reflects current conversion (%id/g, etc.)
* file/xif selection widgets paths are set to last used
0.7.11 2003.06.20
* zh_CN i18n contributed by Wales Wang
* added back in support for internationalization
* ui_common axis works on OS X (was a double free bug, that didn't manifest on linux)
* variance is now calculated correctly (was overconservative previously).
* generalized FIR filtering added
* gaussian filtering done with FIR filtering code, 50% slower, but eliminates
a lot of otherwise redundent code
* median/filter code no longer data type specific.
* ui_series now tracks changes to data sets/rois/and fiducial marks
* fiducial mark "point" parameter eliminated. redundant
* switched over to anti-aliased canvases for all canvases
* misc. fly through canvas fixups
* switched form of raw format entry in .xif files
* rewrote isocontour selecting function, no longer recurseive
* changed look of rendering window/added toolbar
0.7.10 2003.05.27
* enable-win32-hacks option, allows compilation on cygwin/mingw32
* multiple small fixes for cygwin/mingw32
* amitk_dir_sel, slight update, and copied _gtk_fnmatch from gtk
* added slice caching system
* various canvas cleanups
* study can now be active_object -> can rotate entire study from canvas
* default color tables (set in preference dialog) applied to imported data
* enabled subject_weight/injected_dose in medcon_import.c
* resets initial study name on first data set load
* canvas target on/off is now a study parameter, not a preference
* fixes for how multi-frame slices are weighted
* fixed problem with "flashing" canvases -> inadvertent update_layout call
* stereoscopic rendering on inverse background no longer gives gray background
* roi's can now be drawn center-out and edge-to-edge
* smaller font for ui_series
* various ui_series fixes for dealing with no data sets in object list
* can now compile w/o gsl again
* compilation now detects that it's being compiled on OS X
- disables axis guide, errant font handling on OS X was causing crashes
- disables esd, esd was locking files after amide crash
* roi stats are now on a single pane for static studies
* versions of various compiled in libraries shown in about box
0.7.9 2003.04.22
* added subject_weight, injected_dose, and cylinder_factor to data_set object
* can now easily calculate scale factor for %ID/g and SUV
* amitk_object_copy now recurses, children also copied
* dynamic, rotating rendered movies now generated correctly
- was due to a problem with handling of context volume
* series now keeps the correct zoom even if all data sets not shown
* canvas now minimizes thickness correctly if not all data sets shown
* threshold widget more bullet proof, good for retarded data sets
* spin buttons can now accept scientific notation input
* ROI's now calculated correctly for data sets with one of the dim ==1
* fixed filtering (got screwed up in 0.7.7 with add. of scaling_type)
* fixed filtering of dynamic data sets
* fixes for initializing the toolbar
* was not removing the active_object ref. appropriately in amitk_tree
* fixed selection incorrectly going to study row in amitk_tree
* slightly less idiotic handling of cursors
* now requests to change read permissions if import file can't be read
* more intelligent CTI/ECAT data set naming
* more intelligent medcon data set naming
* remove iscontour_3d signal from amitk_canvas by moving code from ui_study_cb.c
* user now asked to specify the depth of geometric ROI's on drawing
* all warnings on reading xml are condensed into a single warning message
* stats now point out difference between voxels and fractional voxels
* variance now calculated using voxels, not fractional voxels
* building rpm/deb files without libecat by default
- xmedcon now supports ECAT 7
* building rpm with enable-amide-debug=no
* fixed help window not popping up problem
0.7.8
* fixed "crazy spin button" bug, again
* stats can now be calculated on only a percentile of the voxels in an ROI
* added principle component analysis to the factor analysis wizard
0.7.7 2003.03.03
* added ability to only look at a subset of the canvases at a time
* added new preference - canvas size can now be fixed to the dimensions of all
data sets, not just the visible ones.
* now have a 3-way linked view too
* view mode is now saved with the study, and is kept with the study object
* fixed problem with amitk_data_set_set_scale_factor not always running when needed
* added scaling_type to amitk_data_set, to indicate dimensionality of scaling array
* ui_time_dialog won't crash if destroyed while event pending
* on data_set_copy min threshold set was getting ignored
* canvas now only uses data sets as bounds. avoids problems with pathologic ROI's
* added .info file generation for fink building
0.7.6 2003.01.18
* fixed long standing bug, dealing with slices generated via nearest neighbhor
over multiple frames. Would cause "flickering" in dynamic movies
* added a factor analysis tool
- implemented Sitek, et al. penalized least squares algorithm
* ui_fly_through's can now be done over time
* ui_series now shows roi's and fiducial marks too
* added an option so that moving of rendered object can be click-n-drag
* encoding of selected objects now in amitk_object
- new signal, "object_selection_changed"
- canvas and tree now automatically adjust for changed in selection
- canvas and tree also keep track of the entire object tree
* more progress bar popups
- histograming
- flythrough and rendering movie progress bars now popup
- cropping
- importing
* dropped the "IDL" file type
* was crashing on interpolation change after object dialog is closed, fixed
* crop dialog: added idle loop handler for canvas update, fixs spin button bug
* scrollbar's on canvas now do a better job of guessing their adjustment values
* managed to get rid of dump double clicking bug on canvas
* study_copy wasn't copying view_start_time
0.7.5 2002.12.17
* medcon importing now allows negative values (had been set to zero)
* added progress bars for rendering context initialization
* added progress bars for series slicing
* added option to set values above threshold to zero when rendering
- a poor man's "strip" function
* projection image voxel units are now in terms of 1/mm^3
* object dialog now lists memory usage of data sets
* fixed unref'ing bug with gtkdial
* compiling without libmdc works again
* slices correctly generated for data sets with negative values
* added ability to leave cross hairs on canvases (in preferences dialog)
* added ability to increase the central "empty area" of cross hairs (in preferences dialog)
* removed annoying characters from suggested export file names
* switched over to using g_try_malloc, g_malloc crashes on failed allocations
* fixed bug in rendering, was allocating 5x more memory then it needed
* allowing optimized versus non-optimized (less memory) rendering
* switched all numerical entries to spin buttons for better behavior on focus out
* rendering is now done in idle loop
* series slicing is now done in idle loop
* ui_series wasn't disconnecting data_set handlers
* added popup dialogs before roi analysis and rendering to allow picking of options
0.7.4 2002.11.26
* switched mpeg encoding to using libfame
* number of voxels now displayed on roi_analysis dialog
* fixed problem with spin_button callback's running on delete_event of program
* better medcon data set name guessing
* threshold doesn't crash now when dealing with negative values
* fixed multiple misuses of EPSILON
* isocontour drawing on dynamic scan frames > 0 caused crashes, fixed
* resizing of 2D isocontour's thickness now works
0.7.3 2002.11.11
* added median filtering to the filter dialog
* distribution data now saved in .xif file when available
* corrected an error in the generation of roi stats
* fixed some small bugs with dual cursor display
* on roi clear, max and min are now recalculated, along with the distribution
* no longer moving canvas view when shifting roi
* now sets thickness correctly when loaded up directly with an imported data set
* added ability to erase data set outside of an roi
* removed slop in moving fiducial marks
* dual cursor display weirdness is now (hopefully) fixed for real
* help is working again
0.7.2 2002.10.20
* anisotropic filtering added, low pass gaussian filter current implemented
* fixed some bugs with rotating/moving data sets
* changed rendering x axis again, think it's right now
* interpolation is now specified per data set
* several entries in amitk_object_dialog are now spinners
* disabled main window resizability, this allows the window to autoshrink
* roi stat's page now comes up as a better size on small screens
* time dialog now resizeable
0.7.1 2002.10.06
* can now generate fly through mpeg movies
* can now choose between blended and overlay fusion modes
* adding a data set cropping wizard
* added the ability to erase a portion of a data set
* most internal formats are now double
* shifting and rotating data sets now uses an overlay image
* can now compile correctly without libgsl support
* alignment's only need 3 points now. problem was with SVD decomposition
* rendering flipped left and right. that's now fixed
* object modification dialog reads in entries a little more sensible now
* big endian raw data importing works again
* speed increase in reslicing. loops now iterate only over portion
of slice that encompases the data set
* fixed some small bugs with changing the view thickness on the canvas
* fixed some small bugs with setting the minimum thickness of a multi data set study
* data set icons now change with the data set's color scale
* added black/white/black and white/black/white colormaps
* cursor now shows current voxel value on active volume
* removed extraneous memory consumption of raw_data_import
* now ignores "per slice" scaling when converting to rendering
0.7.0 2002.8.16
A complete rewrite of almost all of the code in AMIDE
Major Changes:
* ported program to GTK 2.0 and GNOME 2.0
* data set's, roi's, etc. code has now been totally rewritten
- these items are now subclasses of GObject
- any object can be a child of any object
- single object modification dialog that handles all objects
* new .xif format (2.0) that follows new object code
* old .xif format supported via legacy.c
* dropped notion of having a "view coordinate frame", same as base
* canvas updates are now done by an idle handler
* switched internally used data type to double (was float)
Minor Changes:
* when threshold or popup button hit, and dialog already up, brought to front
* dropped all remaining uses of imlib - can now only export images as jpeg...
* user is prompted/warned on inverting an axis
* ROI's can also be used as fiducial markers for alignments.
* added total counts to roi analysis
* single tree now used for linked viewing
* EPSILON value is now used correctly for comparing floats
* ROI rendering conversion speeded up greatly
0.6.4 2002.8.16
* view center notifier on help panel now gives correct location
* fixed subtle floating point rounding error in drawing isocontours
0.6.3 2003.05.27
* Mac OS X doesn't have lround, using rint instead
* canvas arrows weren't being updated with canvas slider change
* with dual cursor display, second display wasn't being adjusted with canvas slider change
* canvas voxel size adjusted when volume's voxel size adjusted
* now better at guessing names from xmedcon imported files
* can now shift data sets around on the canvas
* help info panel slightly rewritten, more legible
0.6.2 2003.05.07
* fixed brown bag type error concerning active volume
0.6.1 2003.05.06
* added a "linked" view mode, so that data sets can be looked at
simultaneously but not superimposed
* spin buttons (zoom, thickness) no longer "go crazy", required a
cludge widget, amitk_spin_button
* when volume/roi/align_pt is changed on canvas, modification
dialog also changes
* when changing layout, canvases now reset coord frame's appropriately
* objects no longer get unselected when ok/apply hit in the
modification dialog
* loading a new study replaces the old study if it's empty
* deleting objects works again
* now confirms quit when unsaved changes exist
* cursors now handled better
* 3D isocontour's now warn if drawn over multiple data frames
* switched all canvases to antialiased
except for amit_canvas and ui_preferences
0.6.0 2003.04.11
* realspace_t's are now a referenced object type
* widgetized the coord frame rotate panel as amitk_coord_frame_edit
* widgetized the canvas as amitk_canvas
* widgetized the tree, got rid of ui_volume, ui_roi, ui_align_pt
* raw_data_read_volume/read_file no longer requires
preinitialized structures
* added "per frame" scaling ability
* added ability to generate movies "over frames"
* added "interpolate between frames" scaling ability.
* changed threshold dialog to a linear layout
* threshold types (per slice/global/etc.) are now for each volume
* setrlimit no longer screws up on MacOS X compile
* threshold widget is no longer so picky about inputs
* fixed some compile/installation problems (thanks to Juha Koivisto)
-unnecessary include of matrix.h (from libecat)
-unintended install of amide_export.h
-libxml include file location changes
0.5.0 2002.02.05
* added ability to place fiducial reference points and align data sets
using a rigid body transform (procrustes method)
* stereoscopic rendering support added
* roi's are now also rendered
* improved alpha blending algorithm
* fixed bug with loading in ASCII data
* imported files now are centered on 0,0,0 to begin with
* added ability to import CTI attenuation and sinogram files
* dropped the use of an inverse axis in the coord frame, as we can just use the transpose
* cheasy eye candy color scales on option menu in amitk_threshold
* removed colons from default exported file names
* roi statistics dialog now is resizeable, scrollable, and lists frame duration
* zoom is now based on all data sets in study, not just currently selected ones
* fixed small configuration bug when compiling without libmdc (thanks Adam Hupp)
0.4.7 2003.12.20
* can now pick between linear and orthogonal layouts for the 3 views
* alignment lines work again
* series views were showing incorrect locations
* can now erase points from an isocontour ROI
* dropped double click activating in tree list, too annoying to support in GDK
* configure script checks for gcc 2.96, and complains loudly if found
0.4.6 2001.12.09
* added 2D and 3D isocontour ROI's
* no more notion of an "active" roi. All roi's are now always moveable
* multiple fixes for closing widgets
* fix for crashing study parameter window
* fixed cti import bug that screwed up the value for small frame durations
* added more logic for dealing with corrupted CTI files
* canvas event processing functions combined and rewritten for clarity
* (X)medcon importing now (hopefully) get's scaling constants right
* specific (X)medcon import fallbacks can be specified by the user
* added ability to load in PDP and VAX raw data formats
* can now handle data set with inf and nan values
* fixed the bug with saving ROI statistics
* gcc requirement is now listed as gcc 3.0.2
0.4.5 2003.11.18
* removed notion of "grains" for calculating ROI statistics
* added a "scaled" color strip to the threshold widget
* redesigned the roi analysis dialog box
* added a button to the roi analysis dialog save stats info to a text file
* fixed minor bug with time button display
* fixed minor bug with threshold changes getting lost in volume dialog
* can now import files from the command line
* dropped 2x2x1 and 2x2x2 interpolations
* gtkgammacurve setting is now faster in ui_rendering_dialog
* can now set the width of drawn roi's as a user preference
* added a gtkdirsel widget, this means .xif directories look more like files
* better (but still not perfect) ROI resizing.
* axis indicators on the roi and data set dialogs
0.4.4 2003.10.31
* rendering dialog changed considerably
- user can now grab canvas and spin it using buttons 1 and 2
- x,y,z rotations have been moved
- all rotations are now relative to the canvas
* rendering parameters dialog fixed up, now allows colormap changing
* rendering movies can now be done over time
* can now change the name of the output movie
* medcon importing importing was a little off, hopefully we
have the right scale factors now
* rewrote the threshold widget as a gtk style widget
* added toolbar and menu to series widget
* allow independent thresholding of series display
* added captions to the series pic's
* now importing correcting the frame durations of dynamic CTI files
* now adding in image offsets from CTI files
0.4.3 2001.10.18
* multiple speed improvements (at least 2x faster now for slice generation)
-changed multiple functions to defines (based on profiling)
-inverse matrix is now pregenerated and stored
-default compile is now -O6
* now have multiple internal data representations (not just float)
* can have multi-dimensional scaling constants, useful for CTI files
and such that have a scaling constant for each plane (let alone frame)
* changed color compositing to an alpha based method, so that
colormaps with white backgrounds can be composited together
* changed voxelpoint_t to 4D, volumes are now considered 4D,
rather than 3D with multiple frames
* new data type, data_set_t. The volume's data set and distribution
array are kept in this format
* interpolation type used for conversion to a rendering context can
now be specified
* can now export renderings to graphics files
* added "reset axis" button to rendering dialog
* fixed a bunch of stupid bugs with the threshold widget
* no longer save max/min in xml files, as these are regenerated
* no longer save far corner in volume .xml files, redundant information
* fixed stupid bug with the raw data import dialog box
* fixed problem with adjusting the dimension of ROI's that were not lined
up with the transverse plane
* wait cursor no longer stays on erroneously when series widget used
* library no longer built by default
0.4.2 2001.10.01
* global scaling is now default for new studies
* threshold arrows now move better with relation to the cursor
* keeps better track of which directory we're current working/saving in
* can now export the views as image files (jpeg/tiff/ps/eps/png/etc.)
* fixed endian issues with loading in floats and doubles
* now ensures save files end with .xif
* numerous build chain bugfixes (now compiles and runs on Mac OS X)
* rewrote trilinear interpolation (much faster)
* dropped bilinear interpolation
(doing it correctly would be slower then trilinear)
* fixed interpolation bug for slices of depth != volume->voxel_size.z
* upgraded almost all the program from using gdk_imlib to gdk_pixbuf
* changed the origin of our image volume to the far left bottom corner
- world coordinate system is now right handed (this is a good thing)
- new .xif version (y axis now points in opposite direction)
- medcon, idl, and pem inputs also updated
- many many updates to make this change work
* elimination of several unneeded and confusing realspace functions
* got rid of the "main" window
* changes with command line parsing, no longer use "-i" to specify the
input file. Also can specify multiple input files
* fixed a long standing problem with destroying already dereferenced
roi_item canvas objects - thanks to awu
* updated config.guess, config.sub, libtool, etc.
0.4.1
* MPEG-1 movie generation capabilities added. Currently, all output
files are saved as "out.mpg", due to an obscure gtk? bug.
* added the ability to do horiz. and vert. leveling with the mouse
* added an ascii file format for the raw data file input function
* added a new menu item, so that users can specify the import method
* added the ability to load in studies from the command line
* you can now compile and run amide as a library, exported functions
are in amide_export.h
* add a creation date to the study data structure
* add 'scan_date to the volume structure
0.4.0
* moved the functionality in the right column to a toolbar on top
* fixed some dumb errors with the mouse-depth change
* couple small documentation updates
* more sensible initial classification function for rendering
* the density and gradient classification functions now work correctly
for multiple volumes
* fixed stupid bug in rendering_context_free()
* rendering parameters dialog changes now make more sense
* did a first-draft of user documentation for rendering
0.3.5
* added Hot Blue/Hot Green colormaps
* changed study tree from GtkCTree to GtkTree
* context sensitive help window working better
* add/edit/delete buttons removed from main panel
* when moving an ROI, the view_center is now linked to the center of the ROI
* disabled bug-reporting, as I only read the sourceforge bug reports
0.3.4
* corrected stupid roi-zooming error, although roi-zooming still not correct
* added a little display to indicate the current coordinates
* added a continously running help window to indicate what the different
mouse buttons indicate at any given time.
* corrected activating an object without selecting it (now not possible)
0.3.3
* added "targets" onto the canvases
* can now change thickness by click-n-dragging on canvases
* fixed rotating of roi's by mouse-click-n-dragging. Zooming still doesn't work right
* mouse cursors now work correctly
* fixed fatal errors in loading in many types of raw data
(unsigned shorts, unsigned ints, etc.)
* fixed a fatal error in referencing a free'd ui_threshold structure
* blank images are now background color, not black
* the currently active object is now noted in the study tree
* code cleanups
0.3.2 2001.06.11
* big endian compile fix
* additional viewing parameters saved in amide study files
0.3.1
* multiple fixes with viewing series of images
* multiple fixes with calculating roi statistics
0.3.0
* added study parameter modification dialog
* accelerated nearest neighbhor reslicing by a factor of 5
* massive bug hunt
0.2.4
* added loading/saving support in an XML file format
0.2.3
* added volume rendering support
* numerous bug fixes
0.1 2001.01.03
* first released version
