05/01/16 -I have optimized the new segmentation algorithm, and it works even better. The improvment is so important (in my mind) that this justifies a major version change: TA is now in v1.5 beta -since segmentation is a slow process I have added an autoclosing dialog that remains open as long as segmentation is taking place 04/01/16 -Drastically improved algorithm and quality of initial segmentation (up to 10 times less errors than previously, but it depends on the original image). Once finalized the new segmentation will be filed online. 27/11/15 -fixed a bug leading to loss of cell contacts in very rare cases (when a cell was englobing all other cells of the image) 19/11/15 -fixed a bug for files affecting files with a several dots (.) in their name/path 17/11/15 -added support for ctrl+click to act as right click for mac users (hope this won't break anything) 20/08/15 -fixed a big cell tracking error introduced during the conversion from packing analyzer to tissue analyser (many thanks to N. Dye and R. Etournay for reporting the bug) 12/01/15 -created an external launcher for TA required to build a FIJI update site for Tissue Analyzer -FIJI update site address is http://sites.imagej.net/TA/ 07/08/14 -renamed a few buttons -move TA parameters from Misc to the lateral toolbar 06/08/14 -updated and improved the launcher 04/08/14 -restored/improved loggers 23/07/14 -added partial support for 96bits RGB images (please report bugs/32 bits measurement errors) 03/07/14 -Ctrl + S should now be Apple/Command + S on mac OS X (not tested yet) 01/07/14 -added a tool to detect and plot rosettes (series of connected bonds all below a certain size) 25/06/14 -added a tool to measure cytoplasmic polarity vector and nematic order 16/05/14 -increased speed of pixel recovery for 48bits images (better implementation but will require more resources) 06/05/14 -improved support for 48bits images -fixing functions using channels to handle 48bits images when possible 05/05/14 -adding support for 48bits images -support for 48bits images required a rename of database column names (*red --> *ch1, *green --> *ch2, *blue --> *ch3) -fixed erroneous GUI parameters reloading -fixed several minor channel related bugs/inconsistencies 28/04/14 -removed a few obsolete buttons to simplify the soft 25/04/14 -reintroduced send to clipboard -fixed minor file name/file extension bugs 16-17/04/14 -added support for dynmaic strecth plot -added support for dynmaic plot of vector polarity (before it was showing nematic polarity instead) -fixed minor list bugs (when list was empty) -fixed drawing of the clone that was not refreshing 01-02/04/14 -fixed out of memory error when using "--> IJ stack" in the viewer tab -added a code to filter out T1 transitions that go back and forth (that are most often caused by segmentation errors when bonds are small) -TA youtube channel is down --> move it to the TA specific channel (mark as private) 19/02/14 -fixed a minor recentering error that prevented tracking to complete properly in some cases 28/10/13 -added a button: "tracking>Convert cell tracks to IJ ROIs" that allows the user to export tracked cells as ROIs in imageJ (the cell id is saved in the name and the slice number as well). These ROIs can be used in combination with the "measure" function of imageJ. 03/10/13 -fixed the remaining bug in the watershed segementation (the bug only affected seeded watershed) -finalized a tool to create bond kymographs -fixed several bugs in the tracking code -finalized a tool to reseed a mask using existing seeds 16/08/13 -added a tool to seed the watershed using a mask (Segmentation>Watershed using mask as seeds) -added support for clones in 'plots>ExportCellData' and 'plots>Export bond data' -added plots for any csv command -added a tool to export any SQL command as a csv file 'plots>ExportSQLCommand' -updated launcher.jar to handle 'tissue_analyzer.jar' 14/08/13 -removed obsolete functions -removed the 'single' selection mode (redundant with 'multi' when only one image is selected) -restored plot cells and plot bonds (plots are now based on the database). In addition it is now possible to decide which columns should be exported and which not and in which order they should be exported 13/08/13 -updated the SQL editor (simplified it and allowed it to deal with the splitted DB) -added a tool to execute javascripts -added a function to recover list from java script -fixed major text (re)loading error in dialogs (text was converted to lower case, causing trouble in some cases) -fixed pointer mispositioning when drift correction was applied (also fixed erroneous mouse coordinates) 12/08/13 -improved the tools to measure cytoplasmic and bond intensity SQLite DB>Measure bond and cytoplasmic inyensity -removed several obsolete classes and dialogs -inactivated buggy translations when the image contains vectorial objects -removed all the jFileChooser in TA and replaced them by the common file chooser 09/08/13 -disabled always on top in the "Vector Graphics Tools" frame to prevent it from hiding the colorchooser -cleaned and simplified database tools -removed obsolete buttons -I now set PRAGMA synchronous = OFF (0) by default for my databases (it increases speed of creation of tables by about 10 to 20 times, impressive, but db may get corrupted if the computer crashes, but no big deal as it can be recreated easily by TA) 08/08/13 -removed several useless buttons -moved a few buttons to more appropriate locations -improved Mask>remove small objects (added a preview) -fixed error with loading of clones in TA -fixed several major bugs associated to plugins -removed obsolete code associated to old plugins -made some minor changes to the layout -fixed minor db bugs -improved and fixed bugs in the "autodetect clone" function -partially fixed Virtual Cloning>detect next row -added a few demo videos 07/08/13 -started to associate buttons to their associated video (the goal is to have one video demo per button) -centralised execution of several buttons in TA -added a tool to manually align images (recenter>Correct alignment) -updated and added a preview for the tool to create a mask (mask>Detect objects (binarise + fill holes)) 06/08/13 -fixed severe mispositioning of the drawing cursor on all systems -fixed image preview using the system viewer when the image is drawn dynamically -fixed glasspane not showing when image is dynamically drawn -improved a bit the display of buttons in viewer -created video demos for several buttons 05/08/13 -changed the size of TA so that it can be used even on a very tiny screen (1024x768 px) -the local watershed can now be executed on several independent cells at the same time (the only requirement is that each of these not-correctly-segmented supercells contains at least two seeds otherwise the cell will simply be ignored). 02/08/13 -files that are not physically produced anymore by TA can be generated on the fly in the viewer, the color coder or the image compositer (and can also be sent to ImageJ) 01/08/13 -removed deprecated raster masks and the code using it (obsolete since the development of ROIs) -the database is better integrated into TA, whenever an action is performed the db should be updated in parallel. The aim is to keep images and the db always coupled, it may fail, especially if the user cancels a running thread for example, not much I can do to prevent that unfortunately -removed several obsolete buttons -removed the possibility to read csv and txt files from within PA (obsolete since the integration of the DB) 31/07/13 -added back the possibility to create a master database containing all the dbs from all frames (made use of the attach and detach functions of sqlite for that) -added vector polarity measurements to the database 29-30/07/13 -mask_recentered and tracked_cells are now obsolete files, the translation is at present dynamically calculated and applied only when really necessary -recentering informations are now coded in the db and translations are calculated on the fly -recoded the "finish all" so that it generates a DB file containing cells and bond infos per image -replaced all calls to "border_cells" images by direct calls to the DB -fixed several bugs associated to the DB -fixed viewer2 that got broken during the transition from PA to TA -removed calls to boundary_plotter2.tif -modified the DB plotter for it to handle the fact that the TA database is not unique anymore -removed several obsolete buttons 26/07/13 -recoded the ROI based watershed so that it is now compatible with biblur and bassin merging, it also works very nicely now (more powerful than the mask based watershed) -removed several other useless buttons from the segmentation tab -I now allow display refresh when watershed is run -fixed a minor channel related bug in "Remove false bonds" -cleaned all the tabs (removed useless buttons + improved tab layout) -created a new frame that will be specifically in charge of detecting and fixing cell tracking errors, most buttons of the tracking panel are being moved there -converted all the cortex tools into a cortex plugin -removed plugins that are not in use anymore 25/07/13 -fixed ans simplified reload the clone in TA -updated the biblur -improved DB individualization -started removing useless image files (obsolete because of the database) -the software now selects the "Segmentation" tab by default at launch (tabs before that include non necessary masking and manual segmentation tools --> should only be used by experts) -removed and regrouped buttons in the segmentation panel -added new functionalities to the "detect bonds" dialog (i.e. bassin merging, small cell removal and single or bi-blur), results are quite impressive as long as fusion parameters are not crazy -I have recoded the watershed in the mask only, it now works like a charm and made it so that it can make use of bassin fusion and biblur 24/07/13 -I thought carefully about the DB and I am now totally convinced that having a common DB for all images analyzed in PA is a very bad idea, therefore each image will have its own DB in the folder that has the same name as the image. Having several DB is safer in cases where users don't always load the same images in the list (or load them in a different order), it is safer for multithreading, it is safer for data loss as not everything is lost if one DB gets corrupted. Having one DB per file also makes it possible to update the DB just for the few files that really need it. DB queries are also made simpler by having one DB per file. It also prevents DB conflicts when working with files in different folders, ... I'll still give the opportunity to create a master DB if the user needs one (I'll append file_name and frame nb to each table to avoid pbs in this merged DB). I have therefore individualized DBs. -fixed several layout bugs 23/07/13 -added a seed prev and a local seed prev function -improved seed prev/next and local seed prev/next (better gestion of channels) -scan area is now a free parameter in local seed next/prev 18/07/13 -I have slightly modified TA so that it can now be run as a FIJI/IJ plugin -fixed a few bugs related to the use of TA as a FIJI plugin -renamed a few files to prevent ImageJ proper to load them to the plugins menu (FIJI was not doing the same mistake) 17/07/13 -I now preserve the position of the scroll bars when another image of the list is selected (if applicable) -added the possibility to rename a file and its associated folder from within TA -fixed a bug that was removing the entire mask when the user was placing one of his seeds on the mask before pressing Ctrl+M 16/07/13 -fixed a bug that prevented proper plugin loading (still need to clean the code though) -polarity is now plotted dynamically from the db (image files are not read anymore), the advantage is the rescaling factor can be tested live and it is now possible to see polarity for each of the channels (for example to see if they differ) 12/07/13 -improved a bit the TA undo function -removed several useless/lengthy .doClick() -fixed a bug reported by a user that was significantly slowing down user drawings (especially visible for big images) 11/07/13 -added a local seeded watersheding function. The idea is that if a bond is missing between two adjacent cells, just click on the center of the two cells that share the undetected bond and then press ctrl+M --> the bond between the two cells will be recreated. Of note, it also works if you have five bonds missing between 5 adjacent cells, in that case just click on the center of the five adjacent cells and then press ctrl+M and the five missing bonds will be plotted. -added the possibility to grab and update just a small region of the mask --> may speed up things a bit 09/07/13 -added a tool 'Remove_false_bonds' to get rid of some of the oversegmentation 03/06/2013 -fixed image loading in TA (force retrocompatibility with previous versions of PA) -fixed drawing of clones/drawing of cell division and death 31/05/2013-03/06/2013 -added support for clone roughness 28/05/2013 -removed local seed next -replaced it by local seed next with bassin fusion -the enhanced watershed algorithm is now fully functional so I decided to replace the old watershed algorithm by the new one throughout the software 27/05/2013 -recoded the seeded version of the watershed -coded a watershed that can merge bassins -cleaned the watershed code 15/03/2013 -fixed a big bug in plugins that prevented them to be loaded 14/03/2013 -fixed a big bug in array sorting (that was not frequently occuring) 28/08/2012 -fixed a huge bug affecting cell tracking (many cells were lost). Many thanks to Raphael for reporting it. -fixed a bug in PA update (bug was on my system only but prevented main jar update on the server) 25/07/2012 -added the possibility to import a LUT from IJ to PA and to apply a palette from PA to an IJ image (panel Misc, button "Import/Apply_PA_LUT_to/from_IJ_image") 20/07/2012 -PA look n Feel (appearance) can now be changed, to do so go to the panel Misc and click on "PA Parameters". I recommend the nimbus L&F, as it is quite ok and cross platform. -fixed a bug in the plugins that prevented some buttons to appear even when the appropriate plugin was there 27/06/2012 -added partial support for FIJI (the path to FIJI should be specified in Misc>PA Parameters) and the FIJI library called fiji-compat.jar (in Fiji.app/jars should be copied to the lib folder of PA) if you do so you would have a fully functionnal FIJI in PA includind 3D Viewer (you can launch it with ctrl+J). Hopefully I will find a way to simplify this procedure in the future -prevented FIJI to close PA when closed. 04/06/2012 -fixed a bug in imageCompositer, the replacement of vectorization tools there prevented "ignore white" from fg image to function properly, this has been fixed 02/06/2012 -added a tool to add ROIs included in other ROIs to the database 28/05/2012 -coupled the partial max projection tool to the ROIs 26/05/2012 -added a what's new in TA window that shows up if there is something new in PA since last startup -centered log window, error window and what's new window on the center of PA main frame -added Zstack position in "quantify ROI" output 24/05/2012 -added support for viewing stacks (thanks to imageJ, see spinner in the bar at the bottom of PA) 22/05/2012 -fixed minor bugs in ROIs -settings for CPU usage (nb of cores PA should use) have been moved to the "PA PArameters" button in the Misc Panel -fixed several bugs in macro mode and allowed more feature to be saved by the macro mode -added plugins functionnality to PA, from now on things might be developed on a plugin basis (--> there will be a core PA that is stable and then there will be plugins developped for some specific applications/users) this will help keep PA simpler and will make it more powerful -added the possibility to dilate ROIs 20/05/2012 -fixed minor ROI bugs -added the possibility to erode ROIs -added a function to extract images from ImageJ -added a shortcut in the ROI panel to recolor a ROI using the last color used (Ctrl + R) 19/05/2012 -new entries in the db (they contain the sum pixel intensity of the cytoplasm) -new function to create a db that contains cytoplasmic intensity of a cell after dilatation of the watershed mask -updated the SQL editor to manipulate tables more easily (EXCEPT commands selects all the columns of a tables except the specified ones; NODUPE selcts all columns except columns that have the same name in the two tables specified) -added the possibility to export SQL table to a tab separated .csv file -added conversion from PA Roi Format ro IJ Roi Format and vice versa -added import from imageJ result/measure window into PA db -added a tool to convert Cells/Bonds to ROI in order to be able to measure them using the imageJ measure tool -added a shortuct for imageJ (Ctrl + J)) -changed the behaviour of the quick list button it now saves depending on the radiobutton selected close to the list -updated the list of PA shortcuts 13/05/2012 -fixed error in conversion of clones to ROI when the clone was touching the border of the image -fixed several minor ROI bugs -sorted list alphabetically in Viewer -added partial support for ROI in viewer 12/05/2012 -added full support for vector transformation of cells and bonds -added ROI quantifications -fixed several minor bugs in ROIs -added transparency support for ROIs -added verbose mode for ROIs -added check for SQLite Db integrity (compared to cell mask) -added a new viewer that is ROI based and can access the database to help users collect data more easily 30-02/04/2012 -cleaned software -added controls for mask validity -added new updaters that are fairly efficient for tracks, masks, ... -fixed drawing issues, things are now drawn exactly at the center of the drawing area before there was a shift of a couple of pixels 29-30/04/2012 -added a tool to convert ROI to clone and vice versa -removed update/progress --> will create a new version of this -fixed the moore algorithm to sort pixels of a contour, it now really is an implementation a real implementation of the moore algo -added the possibility to replace entries in a sql database 25/04/2012 -added broken line ROI -added polygonROI -fixed couple of minor bugs in ROIs 23/04/2012 -finalized quantification of divisions in egg chambers 18/04/2012 -fixed remaining bugs in ROIs -added a tool to create tutorials for PA (in misc, button tutorial mode) 16/04/2012 -fixed bug of windows not showing at the appropriate position with respect to main window -added send to back and bring to front for ROIS -added the possibility to simplify the contour of polygons (useful when contains too many points) 15/04/2012 -finalized ROIs(added alignment, export to SVG, the possibility to create a montage out of images in the list that can be saved as SVG and edited in Illustrator or Inkscape or alike) -added support of grouping of ROIs -added the possibility to swap 2 ROIS/image ROIS -fixed a serialization bug -added Point ROI -added the possibility to edit ROIS 13/04/2012 -added full support for ROIs (ROI Panel) -hairs and Manually drawn divisions are deprecated --> replaced by ROIs (select the type of plot in the ROI Panel combobox) -added temporarily buttons to convert hairs and manually drawn divisions to ROI (but the buttons will be removed in a couple of months) 11/04/2012 -removed jiggling of image around the cursor while drawing -added partial support of ROIs 08/04/2012 -added a control to check that the user is not going to duplicate a track when he clicks on correct track 01/04/2012 -fixed a bug in the "finish all" function (outputed data was ok but the algorithm was slow because of an error in a pixel sorting algorithm) -added a SQLite database to PA -added a plot function that uses the SQLite db to create cells, vectors, nematics, ... plots -added a connection between R and PA to create graphs out using PA SQLite database 11/03/2012 -added an sqlite database to PA (last 3 buttons in plots) 04/03/2012 -cleaned up the saving of PA parameters 01/03/2012 -finalized the new macro mode (please report bugs) 24/01/2012 -added a function to find duplicated cells 23/01/2012 -I have partially multithreaded the "quantify cell divisions 2" tool and forced adjacent divisions with different ids to be treated separately -fixed a big bug in flow calculations 17/01/2012 -added a function to force cell divisions to have the same mask as cell tracks (I highly recommend not to use this function unless you know what you are doing) 15/01/2012 -fixed a bug in the "quantify cell division2" function, that was sometimes not detecting bonds coming from divisions properly. -slightly modified the definition of the velocity measurements 12/01/2012 -added a tool to color code cell division as a function of its stage prophase/metaphase/.... 09/01/2012 -added txt output for flow fields/shear/vorticity and compression 22/12/2011 -changed layout to gain space to be able to put the channel selection tool in "Tools" -moved channel selection to "Tools" (first position) it will affect watershed, autocenter based on 2D correl, polarity,... -in order to gain space, I have moved List related commands such as sorting, ... to right click mouse button popup over the list -in order to gain space, I have moved the macros burron to the right of the list, the search button has been moved as well -fixed a bug in the color selection in "autocenter based on 2D correl" (Daiki) (may create bugs in other places, so please report bugs, if they occur) -added a button in Movies to create a movie out of the current list, the movie will be called list.avi -fixed a bug, that was slowing down the drawing -allow launching of the vectorial editor without having an image in the list 21/12/2011 -added an online updater 19/12/2011 -updated batik libraries (may cause problems) -removed all unnecessary libraries (may cause problems but unlikely)