search

votchallenge / toolkit

The official VOT Challenge evaluation and analysis toolkit
http://www.votchallenge.net/
GNU General Public License v3.0
155 stars 43 forks source link

AR plot - mixed up axis labels #19

Open daniel-d-koenig opened 4 years ago

daniel-d-koenig commented 4 years ago

Hi, I think there is a little typo of the axis labels in AR plot. The following line defines the 'Robustness' and 'Accuracy' labels: https://github.com/votchallenge/vot-toolkit-python/blob/52b5559d74a16518c5970c2f56eacc06cc0a715a/vot/analysis/ar.py#L55

In the following table I have Accuracy = 0.37 and Robustness = 0.01 for evaluation metric values, but in den plot the position of tracker in the AR plot is at around 0.4 for Robustness and 0.1 for Accuracy. I think either the labels or the evaluation results should by changed. image image image

Thanks in advance.

Best regards, Daniel

lukacu commented 4 years ago

Indeed, I think we have fixed this in dev branch, but it got stuck there because I am busy with other projects at the moment.

daniel-d-koenig commented 4 years ago

I installed the current dev branch and generated the graphs, but unfortunately the data is still interchanged. Data on the x axis should be on y axis and vice versa.

Maybe in the following line the indices of data should be changed https://github.com/votchallenge/vot-toolkit-python/blob/458003f27d2f759fd6ada6f237323a1a01e75cd4/vot/document/__init__.py#L76 to handle = self._axes.scatter(data[1], data[0], **style.point_style()) this will lead that 'Accuracy' values are on y axis and 'Robustness' values on x axis.

Another possibility could be to change following function: https://github.com/votchallenge/vot-toolkit-python/blob/458003f27d2f759fd6ada6f237323a1a01e75cd4/vot/analysis/ar.py#L53