Segment Features - Obtain less output

I currently use Segment Features to obtain the features of my segments, logically duhhh ;). But the output contains too many columns for me. I know I can simply use a filter to get rid of these columns but for me it would be nice to configure the node such that it doesn't produce this output, as I plan on analysing large batches of pictures, and I can use all the trimming down on computing time there is. Is there a way to configure the Segment Features node such that it does not compute the following: "Bitmask" and "Source Labeling"?

You can simply turn that off by the disabling the flag "Append Labeling Information". Also, I want to point you to the new Feature Calculation (BETA) node. Maybe this is more convenient to use.

Thanks, that is really a nice node. However I am wondering how it comes that there is ~10% difference in the output for the major and minor axes when using this Feature Calculation BETA node vs the Segement Features node. I wonder what the difference in computation (method) is.

The difference is, that in the new node we first extract a Polygon and then we use the Polygon to calculate everything. As the polygon is defined in real coordinates, differnences between the former implementation and the new node may be present. we work the option to choose whether one wants to calculate the features on a derived polygon or on the pixels itself. Also, we are working to speed-up several of the feature-sets.

Thanks for your quick answer. I however do have a hard time visualising this.

what exactly?

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.