Stroke detection and filtering

As of v0.4.7 this page is substantially out of date. MarkWrite now incorporate sophisticated stroke detection, and handwriting fluency analysis based on maximum velocity peaks. This currently awaits documentation in the meantime please contact the development team with specific questions.

Although it may not make a huge difference, please check what app settings you are using:

In Data Loading Section:

• Filter Imported Data: Tells MarkWrite whether to filter the pen data loaded from a source data file. The default is False. For txyp files you probably want to set this to True as long as the Series detection is working correctly for your files.

• Maximum Series ISI (msec): This could be better labelled, but it allows you to override the default threshold value used by MarkWrite when it parses txyp files into Series. If the threshold value being used creates too many small Series because of false positive series boundaries, stroke segmentation will suffer.

In Stroke Detection:

• All three settings in this settings section are important. Changing the values and reloading a data file will give a sense of whether the stroke parsing is off because of poor default values for these fields, or because of the current algorithm in general.

In general though, I find that the current algorithm is overlay adaptive at detecting local velocity minima, resulting in stroke boundaries being detected because it is a local velocity minima, but the difference between the minima and surrounding velocity values is very small and should really be removed as a stroke boundary point.