Updates history

1.2.5 (2022.11)

• added import of TrackMate tracks;
• added "Show Tracks" option;

1.2.4 (2022.10)

• added Z-range specification for the 3D reconstructions;

1.2.3 (2022.10)

• fixed local maximum calculation to more strict;

1.2.2 (2021.04)

• added "intensity threshold" output to Log file;
• corrected array size (xyOrig) transform allocation for chromatic correction. It would not work/throw exception with images which height is larger than the width (thanks to Klara Jansen/Mithila Burute);
• for 3D (data with Z) color rendering changed maximum possible intensity value to 99% percentile, to account for possible saturated pixels, thanks to Roderick Tas feedback.

1.2.1 (2020.11)

• for chromatic correction/calibration made initial image registration optional (in the initial menu). The reason is that in case of very regular spacing grid it detects false shifts;
• fixed ThunderSTORM import bug, thanks to Oystein Ovrebo feedback.

1.2.0 (2020.08)

• added option to choose a threshold for detection (measured in SNR);
• added preview of detected spots in the detection dialog;
• finished Mavenization of the project;
• added export/import to ONI Nanoimager format and import from ThunderSTORM;
• added fix for "Load large RT", since now FIJI sometimes skips first numbering column;
• original pixel size calculation from Results table during Reconstruction is now rounded to 1 nm;
• changed the way of histogram calculation for detection of image background and noise (on a convoluted image). Instead of 256 bins by default, it used Freedman-Diaconis rule to calculate optimal bin number. It should help with detection on images that contain some large bright spot (nucleus, etc);
• for thresholded spots bigger than typical point size, additional search for local maxima inside is included;
• added Update site;

More feedback from Christophe @cleterrier

• plots sizes displayed for z-calibration are adjusted depending on the screen resolution;
• exported data to ThunderSTORM/Fairy Dust is written using dot as decimal separator independently on the locale;

1.1.6 (2018.04)

• in chromatic correction "warped" channel is MAX projection now, not default;
• error in SNR estimation is fixed;
• "Sort Results" function updated to work with regular ResultsTable (which sometimes has 'hidden' columns);
• Results table precision is now 2 (after floating point), so it should occupy much less space;
• "Load large Results Table" menu item allows to load, well, large Results Table;
• "Load Drift Correction" now automatically checks for the file separator (tab or comma);
• Export/import to MTrackJ is fixed;
• Main menu is rearranged to have two export/import groups;

Thanks to Christophe @cleterrier following updates were made:

• average localization precision is now estimated from detections where localization accuracy is less than 1 pixel (before it was only from true positives and caused crash, especially for 3D data),
• a final number of fitted localizations is now reported to Log file;
• Z coordinate and localization error averaging is added to the 1 px averaging of photoactivation events;
• for drift correction, added an option to calculate drift with first batch of frames as a reference (not consecutive). It should work better with smaller drift;
• added export to ThunderSTORM format;
• added export to Fairy Dust format for web browser based 3D renderer (viewer is available here);

1.1.5 (2016.09)

• error for borders of scaled z colorscale fixed;
• added output of reference points for color correction to results table option;
• drift correction is FFT based now, no limit on maximum displacement between windows;
• error for Y_(px) calculation in 'table' mode for color correction is fixed;

1.1.4 (2016.09)

• added chromatic aberration (color) correction. Needs some testing, when done, gonna be major update;
• zmin and zmax are taken from the Results table, not from stored calibration;

1.1.3 (2016.08)

extensive testing by Christophe @cleterrier lead to:

• fixing 1/2 pixel shift in reporting X and Y localizations;
• error for translation/average photoactivation rendering sequence;
• particle detection with SD ~ 1 pixel efficiency;

1.1.2 (2016.08)

• added calculation of Z-axis localization error;
• for Z-stack reconstruction points are rendered as 3D Gaussians (normalized);

1.1.1 (2016.08)

• added accounting for X and Y "wobbling" in Z calibration;
• added "save", "load" and "show" options for Z calibration;

1.1.0 (2016.08)

• added colorcoded rendering of Z-stacks (max proj and average color);
• it is accompanied by colorbar showing colorcode;
• added multiple LUTS loaded from FIJI/ImageJ;
• added "Scaled colorcode for depth" option that stretches LUTs according to Z particles distribution;
• fixed duplicate fields names to enable call of "Reconstruct Image" from macro;

1.0.2 (2016.07)

• Z-calibration dialog is almost completely re-written;
• parameters as polynomial degree, zero position and range added to dialog and can be changed;
• the dialog now works interactively in "non-modal" mode.

1.0.1 (2016.07)

first addition of 3D astigmatism data processing:

• added dialog for z-calibration creation and fitting;
• added rendering of 3D data that only supports "Z-stack" option.

1.0.0 (2016.07)

first release! many internal and external changes:

• sequence of columns is changed in Results table (reason for version increase), new columns added (errors for SDX,SDY,etc), chi_square is now Rsquared of fit;
• almost all results now in nm units, not pixels! (localization error, SD, etc) check new columns titles;
• the same happened with dialogues, most parameters now in nm (apart from SD of PSF, still in pixels for convenience);
• added detailed and extensive logging for all analysis/reconstruction parameters displayed in ImageJ Log window. Good for keeping tabs;
• "average consecutive localizations" in "Reconstruct" menu now can update results table with new weighted values from averaging;
• final render of "Reconstruct Image" now is in 32-bit mode. Meaning that each particle integrated intensity is equal exactly to 1, i.e. total integrated intensity of image is equal to total number of particles;
• fitting speed is increased even more by adding additional constrain for convergence (error relative increase);
• if ROI is selected before "Detect Particles" start, plugin will report/show only particle inside this ROI;
• Drift correction is now a separate menu item;
• Shift between intermediate reconstructions in Drift Correction now in nm too;
• Drift correction cross-correlation calculation is now normalized and some minor boundary issues are fixed;
• new option to specify pixel size Drift correction intermediate reconstructions (before it was always with average localization precision), it should increase speed of Drift correction;
• in Drift Correction render options of intermediate reconstructions are picked up from the last call of "Reconstruct" function and displayed in parameters window;
• Drift correction results now can by applied to Results table;
• Drift correction results can be saved and later loaded and applied to Results table;
• GPU is temporarily disabled to update for new format Results table later;
• minor bug in calculation of Integrated Intensity is fixed;
• minor bug in Reconstruction is fixed (precision round -> ceil);
• Load Particles and Save Particles functions are re-written for reverse compatibility with previous versions;

0.9.2

• major revision by Jalmar and Jens, implementing GPU support

0.2.4

• added automatic image width and height estimation (max X and Y) in the Reconstruction;
• fixed exception handling during fitting (error with inverse matrix). Some detected spots occasionally were lost because of it;
• fixed some names in the Reconstruction dialog (again). Still this version should be compatible with script;
• in Results table generated after linking I've substituted column 'mark last detections' for 'reverse number of a particle in track' column. It seems to be much more general and universal solution;
• fixed displaying particles at 'Linking'. Now particles could be shown separately without tracks (before it was only together).

0.2.3

• another fix in linking of tracks, now indeed the closest detected particle is used (Eugene);
• another fix in names of fields in "Reconstruction" module (Bas).

0.2.0 (2013.05)

• fixed names of fields and interface of the "Reconstruction" module. Now it is possible to call it correctly from a macro;
• fixed linking of tracks, especially frame's gaps handling;
• added export of tracks to MTrackJ;