General bug fixes and improvements:
- History management has been improved: for databases which have enabled it, data are now linked to the process which have produced them. Such provenance information is available in the GUI (using right click on the eye icons, and in the database browser).
- Using the above history management, database update may be performed incrementally after a distributed execution, which saves a lot of time. It can be used this way as long as the files organization description (ontologies) have not changed.
- Visual improvements: manually modified processes parameters now appear in blue, and locked files appear with a red background.
- New ontology for databases: brainvisa-3.2.0. It does not change files location at all, but interprets them differenty: actually, the "protocol" attribute has been replaced with "center", which is more useful for many users. An existing brainvisa-3.1.0 database can be "upgraded" to brainvisa-3.2.0 just by changing their ontology property in the settings, and then updating the databases. It is recommended to upgrade, because new additions in ontologies may not be done in the older brainvisa-3.1.0.
- make process search match also processes IDs, not only name text.
- GUI improvements to easily manage locked files, and lock/unlock many files at one time: in pipeline nodes popup menus, and in the database browser.
- The database browser can now stop scanning a large tree (hit escape), or refresh selected items (hit F5).
- Don't show Python call stack when a process fails with an error, unless we are in expert user level: such messages are generally meaningless to novice users.
- Reduced the memory needed for database updates.
- (slightly) better management of referentials and databases, which now normally allow to avoid the need of the "check database" process after a distributed execution (of Morphologist, typically).
- Iteration pipelines may now be modified by the user: iteration nodes may be deleted or added after the pipeline is built.
- Several fixes in database management.
- Fixed some memory leaks.
- Processes programmers: a process may now provide a workflow (soma-workflow) for distributed execution, even if the process is not a pipeline or an iteration.
- Facilities are now offered for programming "dynamic pipelines", with varying number of sub-nodes (like iterations). This allows to program some limited kinds of map/reduce pipelines.
- For programmers, the "iteration edition" possibility may also be offered in pipeline nodes which children have all the same type.
General processes:
- A bunch of SPM wrapping processes have been moved in the general Axon processes (in Tools), and allow to wrap normalization, coregistration, and the various tissue segmentation methods. They may be found in more specialized versions (with more specific parameters) in applicative toolboxes.
- New viewer for segmentations comparison.
Morphologist toolbox:
- New version of the Morphologist pipeline: Morphologist 2013. The pipeline offers an improved parallelization for both hemispheres.
- Meshes (hemispheres, G/W interfaces, sulci...) use a different, more accurate, smoothing algorithm, and different decimation settings. As a result, meshes are far more accurate, and also bigger (about 60000 vertices for a G/W hemisphere instead of 30000 formerly).
- New viwer and editor for histogram analysis, allowing comprehensive visual feedback, and interactive modifications.
General bug fixes and improvements:
- When a new version of Brainvisa is available on http://brainvisa.info, a message displayed in Brainvisa documentation panel warns users
- Database history management: it is now possible to store history information about output data generated by Brainvisa processes. History management can be activated for a database through its expert settings activate_history option. History of data can be consulted via the Database browser.
- Invalid processes that are not available in Brainvisa (because an external software is missing for example) are now shown in Brainvisa main window as disabled processes.
- The documentation panel in Brainvisa main window is now a real web browser.
- Reusable Anatomist windows in Brainvisa: Anatomist windows opened by Brainvisa now have an eye button which enables to keep the window opened, even when the corresponding viewer is closed, in order to reuse it for other viewers (and avoid creating new window at various location each time a viewer is started). More information about this new feature in Brainvisa manual.
- New feature in menu View -> Close all viewers: enable to close all windows opened by Brainvisa viewers.
- New selection options in the contextual menu of pipelines in order to unselect steps which would write locked files
- Compressed Nifti and Gifti formats are now default formats
- Optimisations to reduce the access time for big databases.
- Fixes in configuration options: it aborted when a non-existing (or moved away) file was referenced in the options file
- Fix font color in the output panel of Brainvisa processes
Important changes for developers:
- Anatomist in "direct" mode: this change allows more sophisticated viewers (BV/Anatomist combined interfaces)
- Reorganization of Brainvisa python modules: old python modules (neuro*.py) have been moved to brainvisa.processing, brainvisa.data and brainvisa.configuration sub-modules. Some top-level modules (directly used in processes, such as neuroProcesses) are still reachable directly via a proxy module, which will import from the new location (brainvisa.processes for instance). Modules which should not be commonly used from processes (like neuroLog) have been removed from the brainvisa directory.
Morphologist toolbox
- The T1 MRI, Sulci and Morphometry toolboxes are now grouped together into the Morphologist toolbox.
- Snapbase: a new tool to do large scale snapshots of data from Brainvisa databases.
- Fix an incorrect flip in Prepare subject step in a specific orientation.
General bug fixes and improvements:
- Fix a problem with referentials and transformations in Brainvisa databases.
- Fix SPM configuration in Brainvisa preferences.
- Fixes in Soma-workflow integration in Brainvisa.
- Fixes in Morphologist pipeline.
- Fix translations in Anatomist.
General bug fixes and improvements:
- Parallel computing integration: Soma-workflow has now been completely integrated in BrainVisa: processes (pipelines) may be run through a "Run in parallel" button, which allows do dispatch processing across local processors, or on a remote computing resource. A specifig GUI allows to control and monitor all executions. Change the user level to Advanced or more in BrainVISA preferences to use this new feature. More information in the BrainVISA manual.
- GUI changes in BrainVisa main window: the doc panel can now be detached from the main window, or docked at other positions, or hidden. The soma-workflow panel behaves the same, and can be hidden or shown.
- English is now the default language of the graphical user interface.
- Transformations manager performance improvements: it used to drastically slow down BrainVisa, especially when using it the first time (typically when using a viewer on an object with transformations information). It now relies on databasing and is much faster.
- "Reload toolboxes" option in BrainVISA menu: allows to take into account changes in types and ontologies in toolboxes, and detect new processes, without quitting brainvisa.
- Processes and iterations numbering has been revised for a better readability for users.
- New data lock feature: it is now possible to put a lock on data files to prevent BrainVISA processes from overwriting them using the parameter contextual menu in processes. It can be useful when you want to keep a manual modification of a mask for example.
- Use Ipython console shell with ipython >= 0.11.
- Support for SPM8 and SPM8 compiled version (standalone) in config, and in normalization process.
- Better handling of transformations loading in Anatomist when loading objects in anatomist.
- New process InspectMultipleData in data management: allow to easily display viewers/editors icons for a list of data.
- Adapt ffmpeg encoder process to new versions of ffmpeg.
- Fixed environment variables when using external commands from brainvisa.
- Fixes in error handling while reading processes.
- Fixes in file selection of a list of output files.
- Fixes in getting viewers/editors for some data.
- Try better to work cleanly without an environment variable BRAINVISA_SHARE.
- Fixed userLevel of modelgraphinfo viewer.
- Fixed some errors / crashes when quitting brainvisa (by reworking exit code).
- Fix log of Anatomist output in BrainVISA log.
- (developers) Cleanup: removed Qt3 compatibility code.
- T1 MRI toolbox
- More improvements in Morphologist segmentation robustness and accuracy
- Hemisphere and white meshes are now more accurate, prettier, and conform to the spherical topology constraint.
- May use SPM8 and compiled variant to perform normalization.
- SPM and FSL normalization processes can retry after a failure due to a bad initialization (image origin / internal transformation problem).
- Sulci toolbox
- SPAM sulci recognition models data have been reorganized.
- SPAM models viewers have been improved and a bad interaction with other objects in anatomist has been fixed.
General bug fixes and improvements:
- Updated tutorial available
- Fix the field "defined in file" in the Brainvisa ontology documentation.
- Brainvisa Python API documentation in sphinx format available.
- New supported mesh format: MNI .obj
- Fix generic importation of a graph in basic user level mode.
- Fix concurrent creation of a directory in case of parallel executions of Brainvisa.
- New python module to start Brainvisa in batch mode: brainvisa.axon.
In this version, an important change for the developers should be noted : the build tool used to compile from the source code has changed. We gave up our home-made tool build-config and replace it with the standard tool CMake, a cross-platform, open-source build system.
This version also brings a number of bug fixes, and improvements:
- Text search available in the log throught ctrl+F and F3 keys
- Text viewer available for more data types (csv, xml files).
- Brainvisa now uses sip QString API version 2, so it not possible to use QString and QStringList classes anymore, python types should be used instead.
- Fix graphviz packaging which make generation of ontology documentation with types inheritance graphs available.
- Update in programming documentation about GUI customization
- T1 MRI toolbox
- New T1 pipeline named Morphologist. This new version is more robust and efficient.
- T1 Normalization processes (using FSL or SPM) have been moved from fMRI to T1MRI toolbox. So, it is now possible to choose normalization instead of manual localization of AC/PC in the anatomical pipeline without having fMRI toolbox installed.
- Sulci toolbox
- New process SPAM models installation to install SPAM models for sulci recognition. The installation of the required models is checked in the Morphologist pipeline.
- Add new 2011 ANN models to sulci recognition available models
- Bug fix in iterative mode of a sulci graph viewer
- New toolbox Optical Imaging
- BrainRAT toolbox
- Add RGB/RGBA background estimation to subsampling and to intra volume registration process
- Add advanced slice replacement process
General bug fixes and improvements:
- New progress bar in processes and pipeline windows.
- Viewers and Editors are now available for list of data. The default viewer is an iteration of the viewer for the type of data in the list.
- Fix background and foreground color of choice nodes in pipeline graphical interface.
- Fix iteration on Series of 3D volume (Import fMRI data)
- Database browser: enable visualization of unrecognized data if a viewer that accepts the data format exists.
- Fix commands log in windows version.
- Documentation panel : enable links with anchor and open web links in a real web browser when possible.
- Browse button in process iteration dialog always opens the dialog that enables selecting several files at different location on the filesystem. This action was only available through right click before.
- Improvement of database update when a process finishes : intermediate directory items created by the process are inserted in the database; output parameters of a sub-process not in a pipeline are inserted in the database.
- Improvement in Morphometry statistics process : output parameters are now output_directory and output_filename_prefix instead of output_prefix and have default values.
- Improvement in DicomToNiftiConverter process : enable choosing the output filename and giving a directory containing dicom images instead of a file as input parameter.
- New importation process to import a T1 MRI from a Dicom directory using Mricron Dicom converter.
- Fix Single file to serie process : generate the numbers of the serie with a fix number of digit to ensure that the files of the serie are in the correct order.
- T1 MRI : new sulci voronoi step before the folds thickness step, to allow saving and reusing the sulci voronoi maps; New type for lesion mask used in the brain segmentation parameters.
- Sulci : new processes to calculate the mean distance from a lesion mask to sulci.
General bug fixes and improvements:
- New Dicom converter using Mricron. The converter is in toolbox tools -> converters. It is available if Mricron is installed in your computer.
- Fix database conversion process which was not compatible with qt4.
- Fix BrainVISA log file.
- Cortical surface: Fix gyral parcellation (command siParcellation).
- T1 MRI pipeline: test if brain mask is empty or full and fail accordingly
- Sulci: Fix scripts siCSVMapGraph, siDiffModels, siDisplayGraph and annealing_tag which were not compatible with qt4.
- BrainRAT
- Fix issue in intra volume propagative slice registration
- Fix command AimsFFDApply
- Add AimsFlip RGB support
- Add a variation ratio parameter to AimsIndividSlices
- Add RGB support to FlipApplication
- Add threshold parameter to IntraVolumePropagativeSliceRegistration process
- Fix issue in AimsIndividStackSlices and AimsIndividSlices
- First version of BrainVISA using Qt4 API. Be careful, Qt4 API is quite different from Qt3. If some of your own processes use Qt, you may have to port them to Qt4 (See Porting to Qt 4 on Qt documentation pages). This version's source code is still compatible with Qt3, so if you compile from source, you can still use Qt3. Note that the compatibility will not be guaranteed anymore in BrainVISA 4.1.
- New contextual menu on pipelines and iterations enabling easy selection, deselection, opening and documentation display of the different steps.
- Visual feedback in processes graphical interface to show which parameters have been modified.
- BrainVISA documentation panel: keyboard shortcuts for contextual menu items and a new button to reload the current documentation page.
- Attributes sorting enabled in database search interface (available only in Qt4 version).
- Import T1 MRI: the default value for subject attribute is the name of the file.
- Types and formats documentation now shows the hierarchy rules for all available ontologies
- Fix: parameter exactType in WriteDiskItem initialization is now taken into account.
- Sulci toolbox - ergonomy improvements in Sulci recognition pipeline: combo boxes to select the recognition method.
General bug fixes and improvements:
- Several bug fixes with list parameters in processes
- Fix problems with big iterations
- Fix display problems in graphical interface
- New menu to start a python shell
- Several other bug fixes
- T1MRI toolbox
- New step in the T1 pipeline: normalization process using SPM or FSL that can replace manual selection of AC/PC points.
- Fix parameter links problems in T1 pipeline
- Reorganized the grey/white segmentation step of the T1 pipeline since the 2005 variant of it was not doing what it should.
- New importation process for brain masks
BrainVISA package should work on Windows Vista Databases are not automatically updated anymore. It is always asked to the user before. Optimization of databases requests. The results are now shown as a table of attributes values. Easier database request with existing values of databases attributes selectable Databases can be moved without needing update. DiskItem selection window cannot become higher than screen anymore No more useless *.minf created by processes Iteration control : it is possible to stop the current iteration or all iterations. Several other bug fixes
General bug fixes and improvements:
Toolboxes can now completely define new ontologies Some parameters links fixes and improvements Suppressed a latency when clicking the "run" button of a process Fix: some attributes in output disk items browser were erroneously hidden, which actually made it impossible to write some specific data types. Fixed synchonization problem between database content and minf files content, which could cause strange problems when re-processing or re-importing data over existing older ones Several other bug fixes Toolboxes:
Fixes in animation/mpeg processes SPM normalization files conversion has been re-written, and can handle SPM5 and SPM8 files T1: the reorientation routines in PrepareSubject have been re-written. They now handle any orientation T1: new TalairachTransformationFromNormalization process, to import orientation parameters from a normalization file T1: fixes in some memory (re-)allocation functions on Mac, which caused memory consumption to grow anormally and some commands to crash in the T1 pipeline. T1: fixes in 3.1 graphs: on sulcus opening descriptors, which mixed unnormalized CSF volume and normalized sulcus surface, on grey_surface_area, and on cortical thickness which could be inaccurate in some locations (sulci fundi especially) T1: in 3.1 graphs construction or upgrade variants, the process crashed on Windows. We fixed it by disabling multithreading on Windows, and making it optional on other platforms. T1: set total sulci length in global graph attributes of 3.1 sulci graphs Sulci: new multi-subject sulci snapshot process Sulci: new sulci labels transfer process, to transfer a labeling to another sulci graph with a different segmentation and graph structure (but same subject)
General bug fixes and improvements:
Improvement of processes documentation : each data type mentionned in a documentation is a link to a detailled documentation about the data type (inheritance, processes that use it, associated formats, place in Brainvisa databases). The source code of a process can now be displayed in the browser. Bug about process parameters default values. More data types and formats accepted in some processes. Application Qt style can now be changed from BrainVISA configuration panel In databases configuration, ontologies list is not hard-coded anymore Improvements in pipelines remote execution ... and many other fixes and improvements under the scene...
General bug fixes and improvements:
Bug when a BrainVISA database is not available. Bug in database conversion process with sulci directory. Improvement of ListOf parameters management. Possibility to import a T1 MRI in NIFTI format. Improvement of the generic importation process. New process in tools -> documentation to generate documentation about BrainVISA types. ... and many other fixes and improvements under the scene... Toolboxes :
BrainRAT toolbox is now distributed with BrainVISA. The fMRI toolbox will be distributed separately as a downloadable add-on package to provide it its own visibility.
General bug fixes:
More database fixes: especially, optilization issues were fixed, which made acces very slow (especially on a network) and going slower and slower while working in a brainvisa session. The database access interface (database buttons on processes parameters) has been improved, especially for output values: it is now possible to type lists of editable attributes, making it possible to iterate importation processes for instance (iterate to import 3 files, just type 3 subjects names). The "save process" (.bvproc files) feature now also works on iterations. Warning: The structure of bvproc files have changed: new files will not be readable by older versions of Brainvisa. Iteration parameters inputs are a bit smarter and a bit more convivial: parameters with a single value are interpreted as the same value for all iterations. Improvements and customizations in database conversion processes, especially when dealing with unrecognized data. in the T1 pipeline, the newer cortical folds graph creation/upgrade processes have been fixed and optimized: the cortical thickness calculations don't take hours anymore. ... and many other fixes and improvements under the scene...
General bug fixes:
Some database fixes Several other little things...
General bug fixes:
Fixed some importation processes Fixes in database update Fixes in database conversions (3.0 to 3.1) when converting "dirty" databases Fix: in some conditions, user typed values in processes GUI where not taken into account Many less important little fixes
Core:
New organization of BrainVISA processes toolboxes. Each toolbox can define its own processes, extend database ontology, extend BrainVISA GUI (for instance add specific entry in preferences window), etc. A new panel has been added to BrainVISA main window to select toolboxes. BrainVISA user directory ( $HOME/.brainvisaunder Linux and Mac and most oftenC:\Documents and settings\{user name}\under Windows) is now considered as a toolbox named 'My processes'.New preferences panel with a tab for each group of options. Options are now saved in options.minfin BrainVISA user directory.It is now possible to add custom processes trees in the main window (saved in userProcessTrees.minfin BrainVISA user directory).Database system had been rewritten for better performance and scalability. The new system is based on SQLite. Default databases organization had been modified. Old organization (named brainvisa-3.0) is not compatible with newbrainvisa-3.1.0organization. A converter had been written to convert frombrainvisa-3.0tobrainvisa-3.1.0.Added support for NIfTI-1 format. An experimental framework for distribution of processes on several computers had been released. The Python API to control Anatomist from Python had been completely rewritten (and documented). Management of referentials and transformations had been improved. New features in process Database Hierarchy Browser: it is possible to search data, copy data using drag and drop, call a viewer, remove data, and convert graphs (using process Compute Cortical Fold Graph Upgrade From Old) through this interface. When right-clicking a process on the main window, a contextual menu allow to open, edit or iterate the processes. Several BrainVISA can be launched by the same user without messing up the log files. Each BrainVISA instance now have its own session. Processes and their parameters can be saved in a *.bvprocfile. Opening this file in BrainVISA restore the process window.There are so many other things that we truly regret that we did not update this changelog regularly. Toolboxes
T1 MRI:
Tools for segmentation of various anatomical objects (hemispheres, grey matter, white matter, cortical sulci and gyri, etc.) from T1 weighted MRI.Sulci
Diffusion and Tracking:
Diffusion MR image analysis and fiber tracking.
New diffusion model added: QBall imaging. Tracking can be don either with DTI model or QBall model. New regularized probabilistic tracking methods. Datamind:
Cortical surface
Structural analysis
Biology
Nuclear Imaging
Morphometry
TMS
Data management
Core BrainVISA toolbox. Contains processes for databases management: data importation, databases checking, update and exploration.Tools
Core BrainVISA toolbox. Contains general processes that could not be put in another, more specific, toolbox.
General bug fixes:
Fixed a command not found error that prevented to use command lines programs from BrainVISA package (such as Aims... or Vip... commands) when BrainVISA was installed in a directory with space(s) in its name. Diffusion and Fiber tracking:
During fiber tracking, the size of fascicles is limited to 20 000 points to avoid infinite loops in some cases. Improvements
Documentation of a category (i.e. category_documentation.minf file) do not have to be in the main BrainVISA directory to be taken into account.
Bug fixes:
Fixed a crash when using bundles features viewers Log file is not expanded (therefore it is not read) at startup time. It avoids possible errors preventing BrainVISA to start when two BrainVISA instances are writing on the same cache file. Fixed an error in DiskItem comparison operations Changed the way system commands are run by BrainVisa on Mac and Windows. It seems to avoid hangups on Mac when BrainVisa is run from the desk or finder, and allows to stop processes on Windows. The rotating icon animation when processes are running has been made more lightweight: it could cause severe slowdown on old and slow machines. Diffusion and Fiber tracking:
Removed options related to bundles selection from fascicle tracking process An error was fixed in linear interpolation of Aims library. This interpolation is used in several algorithms including diffusion and fiber tracking. Improvements
Link with Matlab: multiple Matlab versions may now be supported. The matlab module in BrainVisa can switch between several pymat module compiled for either matlab 6 or 7. We have built pymat modules for both matlab versions in the binary packages (Linux, Windows, Mac, althrough they have not been tested at all on Mac)
Well, the changelog hasn't been maintained for this version... There has been quite many changes, the most noticeable in the core of BrainVISA is a new pipeline system.
Changes in the core:
Fix in combo boxes creation in database browser that broke compatibility with new Python-Qt binding library (SIP 4). This bug made the importation processes break on the Mac version, and probably the Windows version too. Fix a possible crash in the redraw routines when a database browser combo was selected. Especially, this bug caused dirty crashes on the Mac version that made it quite unusable.
Changes in the core:
Improvements in the installation program: it is not really part of BrainVisa, but there is no other place yet for it. Installation in directories containing space characters now work on all systems (esp. windows), running the SHFJEnvironmentVariables scripts is no longer necessary to run all Aims and Vip commands from a terminal, re-running the setup program is safe and may be used after moving the installation directory. Changes to work with more recent versions of SIP/PyQt Changes in processes:
SPM to AIMS normalization matrixes converter now also works with SPM2 normalization files
Changes in the core:
Fixed an ambiguity in the standard database hierarchy that caused the anatomical pipeline to fail at the end (graphToTalairach) Various fixes in C and C++ libraries used by BrainVisa processes. Amongst other things, this update should fix some problems on Windows XP (ROI graph writing, anatomical pipeline failure when using filenames containing space characters...) GUI improvement in DiskItem finder box for data importation ... Changes in processes:
Fixes in SPM->AIMS transformation converter SPM normalization/coregistration matrix converter, SPM images importation for anatomy, multiple cortical folds recognitions, ... in anatomy/pipeline/PrepareSubject, the parameter allow_flip_initial_MRI has been set to false by default to avoid mistakes Mesh smoothing process failed when filenames contained spaces
Changes in the core:
Setup menus for options Databases setup is done via the menus: editing options.py is no longer necessary. Bug reports options Better help system and french documentation. Ahem, er, yes, french. That's true... Porting to Windows Tooltips in processes GUI A new process category: editors, are similar to viewers but are dedicated to data modification. For now, a "label image" editor calls the ROI edition tool of Anatomist Lots of other things I don't remember Changes in processes:
Diffusion and tracking pipeline Anatomical pipeline:
- subject preparation now takes care of image orientations
- the pipeline also creates the head mesh
Some cleanup has been done in the multiple processes of BrainVISA: the userLevels have been modified in some cases so as to show in basic level only really useful ones (pipelines etc).
There was no changelog at his time.