usage:data_import

Initialisation and data import

Initialisation depends upon the version of FAST installed. For the stand-alone version, simply double click the FAST.exe icon. Alternatively, if you are using the native Matlab version, you can start it by typing:

>> homePanel

into your command line.

In either case, starting FAST will bring up the home panel:

To begin with, all of the modules are inaccessible. As you progress through the processing stages, each module will become available sequentially. To begin, we need to choose a dataset.

To begin processing, click Choose root directory. This will bring up a folder selection dialogue; use this to navigate to your root directory, then click select folder.

This will result in one of two outcomes:

  1. If you have already partially processed the data in this folder, the dataset will already have been converted into FAST's internal format and some processing modules may already have been completed. The home panel will change to allow you to pick up where you left off, making the appropriate modules available. If this is the case, please move onto the tutorials for the relevant modules.
  2. If you have not yet processed this dataset, data import will begin. This converts your dataset to an internal format that minimises processing times, and extracts essential metadata such as pixel size and imaging frequency.

In the case of this second outcome, a second file selection dialogue will appear, which can be used to select the specific file within the root folder you wish to process. Once you have made this choice, a new dialogue box will pop up:

This gives you the option to double the size of your images by upsampling them.

Upsampling does not add new information to your images, but enhances the resolution of the pixel-based segmentation and feature detection stages later on. The trade-offs are increased memory consumption and processing times.
Upsampling is most useful when your objects are small relative to the pixel size. It is usually worth considering if the objects you wish to detect are less than ~15 pixels in width, and is almost essential for accurate segmentation if your objects are less than ~7 pixels in width.

Loading bars indicating the progress of the data import process will now appear. Once it has finished, the Segmentation module should become available:

We see how this works in the next section.

  • usage/data_import.txt
  • Last modified: 2022/09/05 23:03
  • by pseudomoaner