Release Notes
by
Rudolf Höfler
—
last modified
Apr 21, 2016 11:24 AM
Release Notes:
1.6.0 (2015/11/23)
- Switched to PyQt5 as backend
- New Segmentation Plugins
- Using QHelpEngine for online help (F1)
- Index search for keyword
- Table of contents
- Preferences Dialog
- Moved some options from processing pipeline to application preferences
- Path mapping table can be edited directly
- Cluster url and port
- Different application stylesheets
- GUI cleanup
- Only one log window instance
- Removed section "Postprocessing and Plots"
- Menu structure in main window cleaned up
- New icons
- Drag and Drop support to open a settings file.
- Browser - New tab "Event viewer"
- Cluster support
- new version of clusterservice (https://github.com/CellCognition/clusterservice)
- API check between CecogAnalyzer and clusterservice
- Pathmapping: each directory path is checked for existence in the server and client side
- Note: CecogAnalyzer 1.6 and the old versions of the clusterservice are not compatible!
- Data I/O
- Choice between text and hdf output is now exclusive. Text output will be removed in the next release!
- Feature computation
- New Category "Spotfeatures"
- Feature parameters for Haralick and Granulometry
1.5.0 (2014/03/31)
- Error correction:
- R dependency replaced by pure python implementation
- less output files (few big pdfs)
- xml format for contraint-files (incl. schema validation)
- more flexible constraint definitions, masking of single matrix elements
- error correction enabled for tertiary and merged channel
- HMM - choice between "smooting model" (r-replacement) and "Baum-Welch algorithm"
- Eventselection
- release TC3 and unsupervised eventselection
- separate section for tracking and event selection
- new option "event channel"
- Minor GUI improvements, obsolete options removed, ...
- Browser
- option "show by classification" added, to show color code of a predefined classifier
- zoom/pan feature for image view panel
- time slider updates less often
- Documentation updated
- Cluster service moved to a separate repository
- Bug fixes and code cleanup
1.4.1 (2013/10/16)
- fixed battery package
- bug fixes
1.4.0 (2013/04/18)
- added merged channel classifier for feature concatenation of other channels
- plugin system for segmentation
- various bug fixes
- cluster setup simplified
1.2.5 (2012/01/03)
1.2.4 (2011/09/07)
1.2.3 (2011/08/23)
- fixed an issue running on MacOS 10.5
- fixed several GUI bugs and improved usability
- fixed multi-plate support for error correction and enabled error correction features in the GUI
1.2.2 (2011/04/15)
- added support for temporal units (minutes, seconds) for the tracking module in case time-stamp meta-data is available
- added the structure path to the path mapping (for cluster execution)
- change the image path for already scanned structures loaded from pickle files (bug fix)
1.2.1 (2011-04-12)
- changed image import strategy: file info and meta data are loaded only for one plate at a time to reduce memory consumption, changed structure file to Python pickle again
- added progress dialog when plates are changed (higher delay due to dynamic loading)
- fixed overwrite of rendering channels with same name (problem between physical and logical channels)
- browser is reseted after loading of settings and import of image data
- added more warnings and consistency checks for image import
- added optional reformatting of well names (set subwell to 2-digit integer)
1.2.0 (2011/04/05)
- fixed severe crash caused by PyQt4
- browser opens on data import automatically
- added support for multiple plates (including annotation, processing, cluster execution)
- fixed index problem on loading annotations
- fixed channel selection problem
1.1.2 (2011/01/28)
- fixed bug writing CellCounter XML files (time information is based on frame indices again)
- fixed bugs in learning SVM classifier
- added "Show Object Contours" shortcut to the browser
- improved speed for 32bit Windows (VC9 whole program optimization)
1.1.1 (2011/01/24)
- enhanced the browser (look-up tables, faster rendering, support for multiple channels and object regions)
- fixed a bug in setting primary features
- fixed a bug in feature normalization (regularization for constant features)
- changed SVM training to list-use (no zero-insert required)
1.1.0 (2011-01-04)
- redesign of input image scan (save structure file as compressed text in input path)
- added support for gallery images of events (sequence of object images per event and branch)
- added support for full feature export of events
- reactivated cluster support, added "batch size" for experiments with very high number of positions (and no time-lapse)