Published on January 2nd, 2023.
Release 4.1 includes new functionalities and updates to XMapTools and improvements to stability and compatibility.
Table of Contents
- Add tools for single-mask classification and the creation of sub-masks. At the moment, it is recommended to use the Random Algorithm forest for creating sub-masks
- Add a point counting tool to evaluate pseudo-modes; XMapTools selects randomly pixels on the map and estimates the modal abundances (in %). A version with uncertainty estimation using Monte-Carlo can be activated
- Add a functionality to calibrate pixels of a selection of masks for standardisation of EPMA data. Note that merged maps are only generated when the option “Apply to all maps” is selected.
- Add entropic filters to the classification (TEXTF)
- Add the old function to run XMapTools and open immediately a project in the MATLAB version using the command: “XMapTools open ProjectName”
- Add a demo version of Bingo-Antidote 2.0 for test purposes
- Improve stability of XMapTools when a ROI is drawn or edited
- Add a multi-standard calibration option which required several major changes in both the import tool and the calibration module for LA-ICPMS data. This update also includes several improvements and minor corrections. Compatibility with data generated using previous versions is not fully maintained
- Add an option to read the sweep time values for each element from an xml file. Note that the file name should be AcqMethod.xml
- Improve performances for plotting and ROI interaction of the Spider module
Data Visualisation Module
- Improve the global performances of the module by implementing the low ressource mode to all plots. Only 20 % of the data points are plotted when this mode is active; this threshold can be adjusted in the option tab
- Add an option to select a range of data (using the fields Xmin and Xmax) when plotting an histogram
- Add a plotting report generated while plotting a ternary diagram
- Add an option to create a single-mask maskfile from he pixels selected by the Identify
- Add a dropdown menu to select a phase from a maskfile to plot only the pixel compositions of this phase
- Add an option to avoid plotting zeros; this is important for the fitting tools
- Fix several issue in the starting function that would prevent the module to open for special cases (e.g. less than 3 maps)
- Improve the selection of input data for classification
- Add an option to select the method used for data scaling
- Add an option to ensure reproducibility of the classification algorithms based on random numbers (using seeds for the random number generator)
- Add an option to select the number of trees in the random forest algorithm.
- Improve compatibility of the EPMA Converter to import data from JEOL microprobes (Windows)
- Activate map re-sampling and orientation corrections to all data types in the import tools. This resolves an issue when LA-ICPMS data are imported from Iolite. It is not recommended to use these tools for other data types
- Add an option to unfreeze the interface after an error by simply requesting to close the app. This option does not require to have unsaved data as in previous versions
- Add a checking function for the selected training set before classification
- Add a mask file signature to ensure data compatibility throughout XMapTools and add-ons
- Improve interface behavior for button availability, tooltips and error messages
- Improve GUI size for screens with low resolution
- Improve the plot behaviour of the main tree menu: if a category/dataset from Quanti, Merged, Result, Other, the first element in the list is automatically displayed
- Export a file Last_LBCsim.txt in the main directory containing each local bulk compositions simulated by XMapTools for estimating the LBC uncertainty.
- Fix an issue in the classification function when part of the selected maps have been deleted
- Fix a minor issue in the variable initialisation for mask files
- Fix several minor issues in the Converter for LA-ICP-MS; reported by Nils Gies
- Fix several minor issues in the behaviour of ROI for training sets; reported by Nils Gies
- Fix an issue in the project loading function
- Other minor bug and fixes
Was this helpful?
1 / 0