1. Software requirements
The toolbox requires Matlab R2017a or above and Matlab’s image processing and statistics toolboxes. The software can run on Windows (32 and 64 bit), Mac OS X (64 bit), Linx (64 bit) operating systems. The toolbox includes the export_fig toolbox[1] for exporting figures from Matlab to standard image and document formats.
Figure 1. Directory structure of toolbox (b), and software structure overview (a).
Installation of the MWSegEval toolbox is achieved by unpacking the MWSegEval libraries at the user’s choice of directory. The libraries can be obtained as a .zip archive or by cloning the tool repository (http://www.github.com/djkurran/MWSegEval). The toolbox expects the MWSegEval libraries to be present in the directory structure shown in figure 1. A test data set is also provided in the testData folder and is comprised of models represented with square mesh elements and triangular mesh elements that are stored in the SquareMesh_ScatteredDensityBreast and TriangularMesh_HeterogenouslyDenseBreast folders, respectively. Readme files accompany each set of data that explain the data and configuration files.
Figure 2. GUI allows user to set parameters for tasks implemented with functions and classes contained in MWSegEval libraries.
The graphical user interface (GUI) shown in figure 2 may be started by right clicking on MWSegEval_Toolbox.mlapp (figure 1), and executing the ‘Run’ command, or or by right clicking on MWSegEval_Toolbox.mlapp and pressing F9. The GUI allows the user to set parameters and to run the tasks that are implemented with functions and classes contained in the MWSegEval libraries. An overview of the software structure is shown in figure 1. When the user selects a task to run, the GUI reads all of the user parameters and calls function taskManager(). The taskManager() reads the user parameters gathered by the GUI, validates this information, and implements a state machine to run the requested task. The state machine constructs the required objects that call object methods to complete the task and the data structure of the objects hold the results. The state machine carries out error handling and passes error flags and the status of various tasks to the GUI so that the user can monitor the progress of the image analysis. The user requests a sequence of tasks to process and analyze an image. The collective set of tasks comprise a workflow that is described next.