Just like written in the title, do you have any suggestions for threshold distance value for detecting in the wild datasets?? I was wondering if you have a optimum threshold value for overall datasets rather than using thresholds = np.arange(0, 4, 0.01) (mentioned in the code).