Hi!
Congratulations for this great piece of work. We were looking at the results of your script for validation purposes. It seems that the calculation of the aperture irregularity metrics is using the same perimeter as for the edge metrics, namely only one edge for each leaf. Such a definition would not agree with the original paper of Du et al. which considered all edges. In particular, it seems that the definition used in the current algorithm would return numbers smaller than 1 for this metrics.
Best regards
Hi!
Congratulations for this great piece of work. We were looking at the results of your script for validation purposes. It seems that the calculation of the aperture irregularity metrics is using the same perimeter as for the edge metrics, namely only one edge for each leaf. Such a definition would not agree with the original paper of Du et al. which considered all edges. In particular, it seems that the definition used in the current algorithm would return numbers smaller than 1 for this metrics.
Best regards