Clipping Path Length

[lolaBerkowitz]

Hello!
The program seems to be removing 15-20cm from my path lengths consistently. This has resulted in most paths not reaching the platform edge. Do you know what could be causing this?

Many thanks!

[avouros]

Some length of the full trajectories is expected to get lost as some end points are removed when the rodent has reached the platform; this was to avoid unnecessary segmentation when the recording continues dispite that the animal has escaped the maze. Additionaly during the application of the smoothing function (see https://github.com/RodentDataAnalytics/mwm-ml-gen/wiki/Configurations#adv_1) some percentage of length error occurs, which however is extremely small. I will try to overcome this issue on a later release by updating the code implementing the smoothing function.

[lolaBerkowitz]

See the attached pictures comparing the same path plotted in the gui and the second via my own matlab code (same parameters as the gui, except platform is a square). You can see the end is being removed without reaching the platform boundary. The latency gets cut by .5s (5 data points). Most of my trajectories in the gui do not reach the platform like the case here. It seems smoothing occurs after classification, correct? We are noticing this during labeling. Given the below example, would the two explanations you offered above still explain the issue? Thank you!

[avouros]

Yes, this happens because of my first explanation:
"some end points are removed when the rodent has reached the platform; this was to avoid unnecessary segmentation when the recording continues dispite that the animal has escaped the maze. "

Is this causing serious issues? If yes I will will put an ON/OFF switch to this option so that the user can deactivate it.