Expand trajectories part

[henryhetired]

Hi,

Sorry if I miss something completely obvious, but in the last part of the expand trajectories function, when the first if is executed, wouldn't you have already set segment s to be classified meaning the second if would never be executed ?

Should it be only setting the cluster ID to s in the first if and set it to be classified in the second if ?

Thanks a lot !
Henry

[ivansanchezvera]

Sorry mate, I was travelling in vacations and did not have time to check this.
If you don't set the classifying flag to true, the algorithm would never converge.
That is my understanding, Also it is possible that there is an implementation mistake (let me know if you really believe this is an implementation error, so I allocate time to check this issue in greater detail).
If the segments are neighbors, then they should be classified in that cluster given the min number of lines is reached.
As it is the algorithm works and it assigns it segment into a cluster based on density.
Sorry about the messy code, I am now doing some major refactoring.
Best regards,
Ivan