Segmentation using vector mask
We are doing some project work and are using DOF segmented through eCognition. An interesting notion I saw: I wanted to speed up the segmentation process by first doing a vector-based segmentation to limit the area of interest and then segment only the AOI. BUT when I do this, the segmentation time goes up 4 folds and is useless. So I ended up segmenting the whole scene and then applying the vector mask. An interesting thing to notice!
I was sure the first approach would be faster due to limited input from the AOI. :)
Was this article helpful?